%reload_ext autotime
import pandas as pd
import plotly.io as pio
import plotly.express as px # Plotting
from tqdm import tqdm
tqdm.pandas()
import statsmodels.api as sm
import statsmodels.formula.api as smf
from sklearn.ensemble import RandomForestRegressor
pio.renderers.default = 'notebook'
pd.options.plotting.backend = "plotly"
pd.set_option("display.max_columns", None)
pd.set_option("display.max_colwidth", 100)

def parse(number):
    try:
        return float(number)
    except ValueError:
        return None

time: 1.39 s (started: 2025-07-22 14:44:49 +12:00)

df = pd.read_excel("LLM_results.xlsx")
df = df[(df.timeLeft == "closed") & (df.goal > 0)]
df.sentiment = df.sentiment.astype(str).str.split(" ").str[0].astype(float)
df.truth = df.truth.astype(str).str.split(" ").str[0].astype(float)
df.urgency = df.urgency.astype(str).str.split(" ").str[0].astype(float)
df["Target_completion"] = df.amountRaised / df.goal * 100
df

time: 10.2 s (started: 2025-07-22 14:44:50 +12:00)

df.gender.value_counts().head(20)

gender
Female                                                                                2276
Male                                                                                  1648
Other/unknown                                                                          542
Non-binary                                                                               3
Male and Female                                                                          3
Mixed (Male and Female)                                                                  2
Female (target audience is primarily women, but includes non-binary and trans men)       1
Mixed (group of individuals)                                                             1
Female (Lesa), Male (Victor)                                                             1
Other/unknown (as they are twin babies)                                                  1
Other/unknown (animal)                                                                   1
Transgender (Female)                                                                     1
Male (assumed based on image)                                                            1
Female, Male                                                                             1
Transgender (Female to Male)                                                             1
Female (assumed based on context)                                                        1
Other/unknown (nonbinary)                                                                1
Non-Binary                                                                               1
Trans Masc Leaning                                                                       1
Transmasculine                                                                           1
Name: count, dtype: int64

time: 6.86 ms (started: 2025-07-22 14:45:00 +12:00)

print("\n".join(df[df.gender == "Male and Female"].uri.to_list()))

https://givealittle.co.nz/cause/brook-smith-family
https://givealittle.co.nz/cause/cancellation-of-second-street-appeal
https://givealittle.co.nz/cause/help-matt-and-jiwoo-get-out-of-their-rough-patch
time: 4.98 ms (started: 2025-07-22 14:45:00 +12:00)

df.gender = df.gender.str.lower().where(df.gender.str.lower().isin(['male', 'female']), 'other')
df.gender.value_counts()

gender
female    2276
male      1648
other      575
Name: count, dtype: int64

time: 14.9 ms (started: 2025-07-22 14:45:00 +12:00)

df.ethnicity.value_counts().head(20)

ethnicity
Caucasian (assumed based on image)                 1499
European/Caucasian (based on image)                 216
Caucasian (based on image)                          213
European/New Zealander (based on image)             162
Caucasian (assumed from image)                       69
Pacific Islander (based on image)                    69
South Asian (based on image)                         69
Filipino                                             68
South Asian                                          67
Caucasian                                            61
Asian (based on image)                               56
Māori                                                53
European/New Zealander (assumed based on image)      53
European/Caucasian (assumed based on image)          53
Indian                                               49
European/New Zealand European (based on image)       45
Māori (based on visual appearance)                   44
Asian                                                40
European/New Zealand (based on image)                39
Unknown (animal)                                     37
Name: count, dtype: int64

time: 7.58 ms (started: 2025-07-22 14:45:00 +12:00)

df.ethnicity = df.ethnicity.str.extract(r"(.+?)(?= \(|$)")
df.ethnicity.value_counts().head(20)

ethnicity
Caucasian                        1886
European/New Zealander            298
Māori                             289
European/Caucasian                281
South Asian                       184
Asian                             178
Pacific Islander                  128
Unknown                            82
Filipino                           78
European/New Zealand               73
Indian                             69
European/New Zealand European      54
Maori                              47
African                            40
Not specified                      36
Not applicable                     34
Indeterminate/Unknown              33
Mixed                              32
Latin American/Hispanic            29
Māori/Pacific Islander             28
Name: count, dtype: int64

time: 21.7 ms (started: 2025-07-22 14:45:00 +12:00)

df.attractiveness = df.attractiveness.apply(parse)
df[df.attractiveness >= 90]

time: 62 ms (started: 2025-07-22 14:45:00 +12:00)

df.ethnicity = df.ethnicity.replace({
    "European/Caucasian": "Caucasian",
    "European/White": "Caucasian",
    "European/White": "Caucasian",
    "European/New Zealand": "Caucasian",
    "European/New Zealander": "Caucasian",
    "European/New Zealand European": "Caucasian",
    "Māori/Pacific Islander": "Māori",
    "Asian/Pacific Islander": "Pacific Islander",
    "Maori": "Māori"
})
top_20 = df.ethnicity.value_counts().head(20)
display(top_20)
df.ethnicity = df.ethnicity.where(df.ethnicity.isin(top_20.index), "other")

ethnicity
Caucasian                     2594
Māori                          364
South Asian                    184
Asian                          178
Pacific Islander               153
Unknown                         82
Filipino                        78
Indian                          69
African                         40
Not specified                   36
Not applicable                  34
Indeterminate/Unknown           33
Mixed                           32
Latin American/Hispanic         29
Fijian                          24
Indian/Fijian                   24
East Asian                      22
Fijian/Pacific Islander         20
European/New Zealand White      19
Indeterminate                   17
Name: count, dtype: int64

time: 18.7 ms (started: 2025-07-22 14:45:00 +12:00)

df.condition.str.lower().value_counts().head(20)

condition
cancer                     185
breast cancer              143
gender dysphoria           137
stroke                      75
multiple sclerosis          67
none specified              54
endometriosis               45
bowel cancer                39
type 1 diabetes             37
melanoma                    37
mental health issues        36
cerebral palsy              33
terminal cancer             33
infertility                 33
chronic kidney disease      31
obesity                     31
heart attack                29
coronary artery disease     28
lung cancer                 27
ovarian cancer              27
Name: count, dtype: int64

time: 14.3 ms (started: 2025-07-22 14:45:00 +12:00)

df.sentiment.value_counts()

sentiment
70.0     1860
60.0      713
50.0      657
80.0      453
85.0      268
30.0      215
90.0      182
75.0       84
40.0       29
20.0       23
95.0        8
10.0        6
100.0       1
Name: count, dtype: int64

time: 6.56 ms (started: 2025-07-22 14:45:00 +12:00)

df.use.str.lower().value_counts().head(20)

use
medical expenses                                                         2054
medical expenses, travel expenses                                         624
travel expenses                                                           327
medical expenses, travel expenses, lost wages                             306
medical expenses, experimental therapies                                  238
medical expenses, lost wages                                              221
medical expenses, experimental therapies, travel expenses                  93
medical expenses, experimental therapies, travel expenses, lost wages      77
travel expenses, lost wages                                                43
experimental therapies                                                     42
travel expenses, medical expenses                                          31
experimental therapies, travel expenses                                    27
medical expenses, travel expenses, accommodation                           26
funeral expenses                                                           21
lost wages                                                                 20
funeral costs                                                              12
medical expenses, funeral expenses                                          8
educational expenses                                                        6
travel expenses, accommodation                                              6
medical expenses, lost wages, travel expenses                               5
Name: count, dtype: int64

time: 14.9 ms (started: 2025-07-22 14:45:00 +12:00)

df.age_group = df.age_group.str.extract(r"(.+?)(?= \(|$)")
df.age_group = df.age_group.str.lower().where(df.age_group.str.lower().isin(['0-14', '15-64', '65+']), 'indeterminate/unknown')
df.age_group.value_counts()

age_group
indeterminate/unknown    2671
15-64                    1117
0-14                      504
65+                       207
Name: count, dtype: int64

time: 31.3 ms (started: 2025-07-22 14:45:00 +12:00)

df.gender.value_counts()

gender
female    2276
male      1648
other      575
Name: count, dtype: int64

time: 10.6 ms (started: 2025-07-22 14:45:01 +12:00)

df["Charity"] = df.whos_involved.str.contains(r"(Charity)", regex=False)

time: 4.8 ms (started: 2025-07-22 14:45:01 +12:00)

df.to_excel("LLM_results_cleaned.xlsx", index=False)

time: 7.07 s (started: 2025-07-22 14:45:01 +12:00)

df["met_target"] = (df.amountRaised >= df.goal).astype(int)

time: 1.57 ms (started: 2025-07-22 14:45:08 +12:00)

df["parsed_age"] = df.age.apply(parse)
df.parsed_age.value_counts().head(20)

parsed_age
25.0    52
30.0    51
35.0    47
17.0    43
20.0    41
28.0    39
23.0    37
6.0     34
40.0    33
27.0    33
18.0    32
4.0     32
13.0    30
3.0     28
26.0    28
38.0    27
19.0    27
14.0    27
2.0     26
50.0    26
Name: count, dtype: int64

time: 8.81 ms (started: 2025-07-22 14:45:08 +12:00)

df.met_target.value_counts()

met_target
0    3859
1     640
Name: count, dtype: int64

time: 3.98 ms (started: 2025-07-22 14:45:08 +12:00)

df.deservingness = df.deservingness.apply(parse).astype(float)
df.deservingness.describe()

count    4499.000000
mean       91.133585
std         4.859491
min        20.000000
25%        90.000000
50%        90.000000
75%        95.000000
max       100.000000
Name: deservingness, dtype: float64

time: 8.3 ms (started: 2025-07-22 14:45:08 +12:00)

import pandas as pd
import numpy as np
from scipy.stats import f_oneway, chi2_contingency, pearsonr
from sklearn.ensemble import RandomForestRegressor
from sklearn.preprocessing import LabelEncoder

def analyze_feature_influence(df, target_col='met_target'):
    results = {}
    target = df[target_col]
    is_target_numeric = np.issubdtype(target.dtype, np.number)

    for col in df.columns:
        if col == target_col:
            continue
        
        feature = df[col]
        if feature.isnull().any() or target.isnull().any():
            continue  # skip missing for simplicity

        if is_target_numeric:
            # Target is numeric
            if np.issubdtype(feature.dtype, np.number):
                # Pearson correlation
                corr, p_value = pearsonr(feature, target)
                results[col] = {
                    'test': 'pearsonr',
                    'stat': corr,
                    'p_value': p_value
                }
            else:
                # ANOVA (feature is categorical)
                groups = [target[feature == val] for val in feature.unique()]
                stat, p_value = f_oneway(*groups)
                results[col] = {
                    'test': 'anova',
                    'stat': stat,
                    'p_value': p_value
                }
        else:
            # Target is categorical
            if not np.issubdtype(feature.dtype, np.number):
                # Chi-squared test
                contingency = pd.crosstab(df[col], target)
                stat, p, _, _ = chi2_contingency(contingency)
                results[col] = {
                    'test': 'chi2',
                    'stat': stat,
                    'p_value': p
                }

    return pd.DataFrame(results).T.sort_values('p_value')

# Usage:
# df = pd.read_csv('your_data.csv')
df.met_target = df.met_target.astype(bool)
analyze_feature_influence(df, target_col='met_target')

time: 679 ms (started: 2025-07-22 14:45:08 +12:00)

df.smiling = df.smiling.astype(bool)

time: 854 μs (started: 2025-07-22 14:45:08 +12:00)

df["have_age"] = ~df.parsed_age.isna()
df[["met_target", "have_age"]].value_counts()

met_target  have_age
False       False       2509
            True        1350
True        False        440
            True         200
Name: count, dtype: int64

time: 9.29 ms (started: 2025-07-22 14:45:08 +12:00)

df["log_amountRaised"] = np.log1p(df.amountRaised)
df["log_amountRaised"].describe()

count    4499.000000
mean        7.422533
std         1.885026
min         0.000000
25%         6.208588
50%         7.476472
75%         8.764990
max        13.388103
Name: log_amountRaised, dtype: float64

time: 7.58 ms (started: 2025-07-22 14:45:08 +12:00)

df.start = pd.to_datetime(df.start, format="mixed")
df.end = pd.to_datetime(df.end, format="mixed")
df["days"] = (df.end - df.start).dt.days
df.days.describe()

count    4499.000000
mean      131.302734
std       141.845040
min         0.000000
25%        32.000000
50%        90.000000
75%       182.000000
max      1755.000000
Name: days, dtype: float64

time: 274 ms (started: 2025-07-22 14:45:08 +12:00)

formula = "log_amountRaised ~ days + sentiment + smiling + gender + truth + urgency + parsed_age"
logit = smf.glm(formula=formula, data=df, family=sm.families.Gamma(link=sm.families.links.Log())).fit()
display(logit.summary())
display(f"AIC (Logit): {logit.aic:.2f}")
probit = smf.glm(formula=formula, data=df, family=sm.families.Gamma(link=sm.families.links.Log())).fit()
display(probit.summary())
display(f"AIC (Probit): {probit.aic:.2f}")

/home/ubuntu/givealittle/.venv/lib/python3.10/site-packages/statsmodels/genmod/families/family.py:812: RuntimeWarning:

divide by zero encountered in log

/home/ubuntu/givealittle/.venv/lib/python3.10/site-packages/statsmodels/genmod/generalized_linear_model.py:1891: RuntimeWarning:

invalid value encountered in scalar subtract

'AIC (Logit): -inf'

/home/ubuntu/givealittle/.venv/lib/python3.10/site-packages/statsmodels/genmod/families/family.py:812: RuntimeWarning:

divide by zero encountered in log

/home/ubuntu/givealittle/.venv/lib/python3.10/site-packages/statsmodels/genmod/generalized_linear_model.py:1891: RuntimeWarning:

invalid value encountered in scalar subtract

'AIC (Probit): -inf'

time: 118 ms (started: 2025-07-22 14:45:09 +12:00)

df.met_target = df.met_target.astype(int)
smf.ols(formula="Target_completion ~ days + sentiment + smiling + gender + truth + urgency + parsed_age", data=df).fit().summary()

time: 48 ms (started: 2025-07-22 14:45:09 +12:00)

import seaborn as sns
sns.regplot(x="amountRaised", y="smiling", data=df, logistic=True)

<Axes: xlabel='amountRaised', ylabel='smiling'>

time: 5.88 s (started: 2025-07-22 14:45:09 +12:00)

smf.ols(formula=formula, data=df).fit().summary()

time: 63.2 ms (started: 2025-07-22 14:45:15 +12:00)

model = RandomForestRegressor()
cols = "sentiment + smiling + gender + truth + urgency + ethnicity".split(" + ")
X = pd.get_dummies(df[cols])
y = df["met_target"]
model.fit(X, y)

RandomForestRegressor()

time: 572 ms (started: 2025-07-22 14:45:15 +12:00)

list(zip(cols, model.feature_importances_))

[('sentiment', np.float64(0.20735737033894144)),
 ('smiling', np.float64(0.11204050356227582)),
 ('gender', np.float64(0.09140590737411987)),
 ('truth', np.float64(0.215663147060733)),
 ('urgency', np.float64(0.049351978324802946)),
 ('ethnicity', np.float64(0.04748830954506516))]

time: 16.6 ms (started: 2025-07-22 14:45:16 +12:00)

df.region.value_counts().head(20)

region
Auckland                1331
Canterbury               555
Wellington               489
Waikato                  385
Bay of Plenty            293
Nationwide               210
Otago                    168
Northland                161
Hawke's Bay              134
International            107
Taranaki                 100
Southland                 71
Manawatu-Wanganui         63
Nelson / Tasman           62
Manawatu / Whanganui      51
Manawatu-Whanganui        51
Marlborough               37
Gisborne                  34
Nelson                    28
West Coast                24
Name: count, dtype: int64

time: 8.73 ms (started: 2025-07-22 14:45:16 +12:00)

	Index	title	pitch	hero	uri	amountRaised	goal	timeLeft	action	actionUri	start	end	n_questions	location	n_donors	description	use_of_funds	whos_involved	moderated	n_updates	updates	_21	text	condition	ICD10	ICD	name	gender	age	age_group	ethnicity	urgency	sentiment	truth	notes	smiling	deservingness	attractiveness	use	region	narrative_clarity	narrative_quality	emotional_tone	image_type	face_visible	facial_expression	image_quality	Target_completion
31	31	Helping a person in need	Helping out a person in need that would drop his hat to help someone else	https://static.givealittle.co.nz/assets/hero/ed760e80-7e10-4078-b7a8-b2ec0104419b-320:False	https://givealittle.co.nz/cause/helping-a-person-in-need	1700.0	2000.0	closed	Read more	https://givealittle.co.nz/cause/helping-a-person-in-need	29 May 2025	8 Jun 2025	0	Blenheim, Marlborough	54.0	Tok had a serious injury while running — he tore a muscle, which led to internal bleeding and ev...	medical care where needed and general stuff around the house if needed also and utility's	Created by, and paying to a verified bank account of,\r\nMatthew bergman on behalf of Tok H...	Page Moderated	3.0	Big thanks\n7 June 2025\nHey team huge thanks to all that have donated Tok is humbled beyond wor...	NaN	Title: Helping a person in need\nPitch: Helping out a person in need that would drop his hat to ...	Muscle tear leading to internal bleeding and infection	S86.0	Chapter S00-T88 - Injury, poisoning and certain other consequences of external causes	Tok	Male	Indeterminate	indeterminate/unknown	Māori (based on visual appearance)	80.0	70.0	90.0	Age was not explicitly mentioned in the text, so it is marked as indeterminate. Ethnicity was in...	True	90	60	medical expenses, travel expenses	Marlborough	80	75	grateful	portrait	True	smiling	medium	85.000000
56	56	Please help Walter 🙏 to get the veterinary care he needs.	Please help us 🙏 Walty needs bloodwork to rule out cancer 😪	https://static.givealittle.co.nz/assets/hero/a5763829-1835-4780-adc3-b2e500bd5557-320:False	https://givealittle.co.nz/cause/please-help-walter-to-get-the-veterinary-care-he	380.0	500.0	closed	Read more	https://givealittle.co.nz/cause/please-help-walter-to-get-the-veterinary-care-he	22 May 2025	1 Jun 2025	0	Otago	14.0	Walter has been my life for 15 years, since he was a pup. It has always being just the two of us...	The money raised will go into Walts vet costsFull bloods, medication etc at the moment is quoted...	Created by, and paying to a verified bank account of,\r\nKerri Back	Page Moderated	1.0	Bloodwork done.\n23 May 2025\nThank you so much for your kind donations.\nWalty has had his bloo...	NaN	Title: Please help Walter 🙏 to get the veterinary care he needs.\nPitch: Please help us 🙏 Walty ...	Pancreatitis	K85	Chapter K00-K95 - Diseases of the digestive system	Walter	Other/unknown	indeterminate/unknown	indeterminate/unknown	Unknown (animal)	80.0	70.0	90.0	The subject is a pet dog, not a human. Age and age group cannot be determined precisely.	True	90	80	medical expenses	Otago	85	80	grateful	portrait	True	smiling	high	76.000000
59	59	Medical Treatment for Atish Lal for Coronary Artery Disease	medical Treatment	https://static.givealittle.co.nz/assets/hero/0a81d0ab-2bfc-4890-8049-b2e3017fea04-320:False	https://givealittle.co.nz/cause/medical-treatment-for-atish-lal-for-coronary	380.0	14900.0	closed	Read more	https://givealittle.co.nz/cause/medical-treatment-for-atish-lal-for-coronary	20 May 2025	6 Jun 2025	0	Auckland	6.0	Dear Family and Friends,\nI hope this message finds you well.\nI am reaching out today with a hu...	Estimated costs for Treatment & Travel:- PTCA treatment package + home stay : $11,000 USD- Reser...	Created by\r\nShaneel Prasad \n\n\n\n\r\n Paying to a verified bank account of\r\nPriya Go...	Page Moderated	NaN	NaN	NaN	Title: Medical Treatment for Atish Lal for Coronary Artery Disease\nPitch: medical Treatment\nDe...	Coronary Artery Disease	I25.1	Chapter I00-I99 - Diseases of the circulatory system	Atish Lal	Male	Not specified	indeterminate/unknown	Indian (assumed based on name and context)	90.0	70.0	95.0	Age was not explicitly mentioned in the text; ethnicity inferred from name.	False	90	50	medical expenses	Auckland	80	75	grateful	portrait	True	neutral	medium	2.550336
74	74	Help me to reclaim my freedom.	Help my to reclaim my freedom and contribute to my community.	https://static.givealittle.co.nz/assets/hero/0e91c689-ef0e-4c5c-aa9d-b2e0010f721b-320:False	https://givealittle.co.nz/cause/help-me-to-reclaim-my-freedom	2999.0	6000.0	closed	Read more	https://givealittle.co.nz/cause/help-me-to-reclaim-my-freedom	17 May 2025	5 Jun 2025	0	Bay of Plenty	15.0	Two years after my first Covid infection, I’m still living with debilitating Long Covid, Chronic...	The money will be spend on a mobility scooter that can be folded down and taken in the car so I ...	Created by, and paying to a verified bank account of,\r\nAmy Cole	Page Moderated	NaN	NaN	NaN	Title: Help me to reclaim my freedom.\nPitch: Help my to reclaim my freedom and contribute to my...	Long Covid, Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (ME/CFS), Post-exertional Malaise...	G93.3 (Postviral fatigue syndrome)	Chapter G00-G99 - Diseases of the nervous system	Amy	Female	Indeterminate	indeterminate/unknown	Caucasian (assumed based on image)	80.0	70.0	90.0	Age was not explicitly mentioned in the text, so it is marked as indeterminate. Ethnicity was in...	True	90	70	medical expenses	Bay of Plenty	85	80	grateful	portrait	True	smiling	high	49.983333
75	75	Help My Mum Fight Cancer - Every Bit Counts	Rasing funds for my mum's cancer treatment in India - every donation brings us closer to saving ...	https://static.givealittle.co.nz/assets/hero/3b1e6715-fdbd-4e4c-b787-b2e20160c6d0-320:False	https://givealittle.co.nz/cause/help-my-mum-fight-cancer-every-bit-counts	630.0	50000.0	closed	Read more	https://givealittle.co.nz/cause/help-my-mum-fight-cancer-every-bit-counts	17 May 2025	19 May 2025	0	Otahuhu, Auckland	26.0	I'm reaching out with a heavy heart and hope in my hands. My mum, the strongest person I know, h...	Hospital bills, chemotherapy, travel expenses, medications, post-treatment care in India.	Created by, and paying to a verified bank account of,\r\nAlana Al Jebin on behalf of My mot...	Page Moderated	NaN	NaN	NaN	Title: Help My Mum Fight Cancer - Every Bit Counts\nPitch: Rasing funds for my mum's cancer trea...	Cancer	C00-C97	Chapter C00-D49 - Neoplasms	Not explicitly mentioned	Female	Indeterminate	indeterminate/unknown	South Asian (based on name and context)	90.0	70.0	95.0	Age and specific ethnic details not provided in text; inferred ethnicity based on name.	False	95	50	medical expenses, travel expenses, medications, post-treatment care	Auckland	85	80	hopeful	portrait	True	neutral	medium	1.260000
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
11184	11192	New Zealand MindBody Trust web-based 'whole person' resources for healthworkers and their patien...	The NZ MindBody Trust web-based project to support and educate clinicians and patients in a 'who...	https://static.givealittle.co.nz/assets/hero/158d0e80-2c47-449a-a05e-a603014e786e-320:False	https://givealittle.co.nz/cause/mindbodyresourcesforpatientsandhealthworkers	6700.0	9000.0	closed	Read more	https://givealittle.co.nz/cause/mindbodyresourcesforpatientsandhealthworkers	9 May 2016	15 Jul 2016	0	Auckland	36.0	The NZ MindBody Trust (2004) is concerned with 'whole person' and person-centred approaches to h...	NaN	Created by, and paying to a verified bank account of,_x000D_\n _x000D_\n New Zealand...	Page Moderated	2.0	The Give-A-Little Campaign Result\n12 July 2016\nThe MindBody Trustees want to thank you all for...	NaN	Title: New Zealand MindBody Trust web-based 'whole person' resources for healthworkers and their...	Chronic illness	R53	Chapter R00-R99 - Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere c...	Not specified	Other/unknown	Indeterminate	indeterminate/unknown	Not specified in text, inferred as likely New Zealand general population based on context	70.0	80.0	90.0	The text does not specify a particular individual, so details like name, gender, age, and ethnic...	False	85	50	medical expenses, experimental therapies, travel expenses, lost wages	Auckland	90	85	hopeful	symbolic	False	not_detectable	high	74.444444
11186	11194	Diane's cyst removal to continue my dream job	Large Radial Ganglion Cyst Removal	https://static.givealittle.co.nz/assets/hero/06af4a19-e239-44bb-a5a8-a60400ab4982-320:False	https://givealittle.co.nz/cause/wristoperation	730.0	5000.0	closed	Read more	https://givealittle.co.nz/cause/wristoperation	9 May 2016	3 Jun 2016	0	Northland	11.0	I am in need of a large ganglion cyst removed from my left wrist. It has entwined itself with 2 ...	NaN	Created by, and paying to a verified bank account of,_x000D_\n _x000D_\n Diane Crocker	Page Moderated	NaN	NaN	NaN	Title: Diane's cyst removal to continue my dream job\nPitch: Large Radial Ganglion Cyst Removal...	Ganglion Cyst	M71.0	Chapter M00-M99 - Diseases of the musculoskeletal system and connective tissue	Diane	Female	Indeterminate	indeterminate/unknown	Caucasian (assumed based on image)	80.0	50.0	90.0	Age was not specified in the text; ethnicity was inferred from the image.	False	85	60	medical expenses	Northland	70	65	desperate	portrait	True	neutral	medium	14.600000
11194	11202	A thank you to Waikato NICU for helping so many little babies and their families	Raising funds for Waikato NICU, and a portable cardiac monitor to support the miracle babies.	https://static.givealittle.co.nz/assets/hero/75b48fa1-0752-44b4-8484-a5fb015ecab2-320:False	https://givealittle.co.nz/cause/charlieandoliver	8912.7	35000.0	closed	Read more	https://givealittle.co.nz/cause/charlieandoliver	3 May 2016	30 Jun 2017	0	Waikato	100.0	This give-a-little page aims to raise funds to allow the NICU at Waikato Hospital to purchase a ...	NaN	Created by_x000D_\n _x000D_\n Rebecca Clews_x000D_\n \n\n\n\n\n\n\n_x000D_\n ...	Page Moderated	4.0	Quiz Night!\n7 November 2016\nGot no plans on the 17th of November? Come on down to the Speights...	NaN	Title: A thank you to Waikato NICU for helping so many little babies and their families\nPitch: ...	Premature birth	P07	Chapter P00-P96 - Certain conditions originating in the perinatal period	Charlie and Oliver	Male	Newborn (preterm)	0-14	Asian (based on image)	90.0	80.0	95.0	The text describes a campaign for premature twins, focusing on the need for a portable cardiac m...	True	95	70	medical expenses	Waikato	90	85	grateful	portrait	True	smiling	medium	25.464857
11195	11203	Help restore David’s smile after being assaulted	Talented professional musician and entertainer David Shanhun needs your help to restore his teet...	https://static.givealittle.co.nz/assets/hero/ee0d17f1-b6a5-4d41-85e9-a5fa0166b741-320:False	https://givealittle.co.nz/cause/help-restore-davids-teeth-and-smile	8048.0	54500.0	closed	Read more	https://givealittle.co.nz/cause/help-restore-davids-teeth-and-smile	3 May 2016	31 Jul 2016	0	Auckland	115.0	David was assaulted on the 16th of April 2016. He had 2 teeth knocked out completely, and 10 tee...	NaN	Created by_x000D_\n _x000D_\n Fiona Shanhun_x000D_\n \n\n\n\n\n\n\n_x000D_\n ...	Page Moderated	2.0	Health/Teeth/Face Update\n24 May 2016\n*Message from David:*\nThank you all so much for your...	NaN	Title: Help restore David’s smile after being assaulted\nPitch: Talented professional musician a...	Dental Trauma	S02.5	Chapter S00-T88 - Injury, poisoning and certain other consequences of external causes	David Shanhun	Male	Indeterminate/Unknown	Indeterminate/Unknown	Caucasian (assumed based on image)	80.0	70.0	90.0	Age and ethnicity inferred; no explicit details provided in text.	True	90	75	medical expenses	Auckland	85	80	grateful	portrait	True	smiling	high	14.766972
11204	11212	TRUS Biopsy Machine for Taranaki	The men of Taranaki need this TRUS Biopsy Machine to diagnose Prostate Cancer. Reduce waiting t...	https://static.givealittle.co.nz/assets/hero/c9169845-63b5-4c1e-bb50-a5fc00e5abd0-320:False	https://givealittle.co.nz/cause/prostatebiopsymachinenewplymouth	1910.0	20000.0	closed	Read more	https://givealittle.co.nz/cause/prostatebiopsymachinenewplymouth	29 Apr 2016	31 Aug 2016	0	Taranaki	15.0	Currently the TRUS (Trans Rectal Ultrasound) machine for diagnosing prostate cancer is only avai...	NaN	Created by, and paying to a verified bank account of,_x000D_\n _x000D_\n Rotary Club...	Page Moderated	NaN	NaN	NaN	Title: TRUS Biopsy Machine for Taranaki\nPitch: The men of Taranaki need this TRUS Biopsy Machin...	Prostate Cancer	C61	Chapter C00-D49 - Neoplasms	Not specified	Male	Indeterminate	indeterminate/unknown	Not specified in text; assumed to be New Zealand European/Pākehā based on context	80.0	70.0	90.0	The text does not specify a specific individual, so details like name, age, and ethnicity are in...	False	90	50	medical expenses	Taranaki	80	75	hopeful	symbolic	False	not_detectable	high	9.550000

	Index	title	pitch	hero	uri	amountRaised	goal	timeLeft	action	actionUri	start	end	location	n_donors	description	use_of_funds	whos_involved	moderated	n_updates	updates	_21	text	condition	ICD10	ICD	name	gender	age	age_group	ethnicity	urgency	sentiment	truth	notes	smiling	deservingness	attractiveness	use	region	narrative_clarity	narrative_quality	emotional_tone	image_type	face_visible	facial_expression	image_quality	Target_completion
169	169	A Helmet for Amelia’s Bright Future	Helping baby Amelia get the orthotic helmet she needs — because every little head deserves a hea...	https://static.givealittle.co.nz/assets/hero/1767aff8-a54d-40d8-9566-b2c7017108a3-320:False	https://givealittle.co.nz/cause/a-helmet-for-amelias-bright-future	2501.0	4200.0	closed	Read more	https://givealittle.co.nz/cause/a-helmet-for-amelias-bright-future	22 Apr 2025	13 May 2025	Hamilton, Waikato	75.0	My bright baby girl, Amelia, has been diagnosed with cranial deformities. The recommended treatm...	Orthotic cranial helmet and therapy for Amelia at Orthotic House, NZ (https://www.orthotichouse....	Created by, and paying to a verified bank account of,\r\n \r\n Dil Askarova\r\n ...	Page Moderated	NaN	NaN	NaN	Title: A Helmet for Amelia’s Bright Future\nPitch: Helping baby Amelia get the orthotic helmet s...	cranial deformities	Q75.3	Chapter Q00-Q99 - Congenital malformations, deformations and chromosomal abnormalities	Amelia	female	infant	0-14	Asian	90.0	85.0	95.0	No specific warnings or errors detected.	True	95	90.0	medical expenses	Waikato	90	85	hopeful	portrait	True	smiling	high	59.547619
3116	3118	Isla’s Hip dysplasia journey - pavlik harness	Little Isla and her hip dysplasia journey 💕	https://static.givealittle.co.nz/assets/hero/799380ba-e48c-4051-9e59-af9a016d59a4-320:False	https://givealittle.co.nz/cause/islas-hip-dysplasia-journey-pavlik-harness	92.0	800.0	closed	Read more	https://givealittle.co.nz/cause/islas-hip-dysplasia-journey-pavlik-harness	30 Jan 2023	15 Feb 2023	Otago	7.0	Isla is 11 weeks old and has just been diagnosed with hip dysplasia.\nShe is now in a pavlik har...	Clothing, Car seat, fuel to get to and from specialist appointments.	Created by, and paying to a verified bank account of,_x000D_\nKerra McLellan	Page Moderated	1.0	Islas journey\n14 February 2023\nWe have been gifted some clothing!\nShare this update\n0 commen...	NaN	Title: Isla’s Hip dysplasia journey - pavlik harness\nPitch: Little Isla and her hip dysplasia j...	Hip dysplasia	M00-M99	Chapter M00-M99 - Diseases of the musculoskeletal system and connective tissue	Isla	female	11	0-14	Caucasian	80.0	60.0	90.0	No explicit mention of ethnicity in text; inferred from image.	True	95	90.0	medical expenses, travel expenses	Otago	80	70	hopeful	portrait	True	smiling	high	11.500000
6455	6459	10for10	We want you to donate $10 for the 10% of babies by doing 10 days of 10 exercises and then nomina...	https://static.givealittle.co.nz/assets/hero/4227fe37-03cb-4c72-ada5-ab9e00f2557c-320:False	https://givealittle.co.nz/cause/10for10	2325.0	5000.0	closed	Read more	https://givealittle.co.nz/cause/10for10	15 Apr 2020	15 Jul 2020	Nationwide	102.0	Each year 10% of babies are born prematurely or at full-term with significant ongoing medical is...	These funds will help us supply Essential Packs for families, Winter Warmer packs for vulnerable...	Created by The Little Miracles Trust, paying to a verified bank account of_x000D_\n _x000D_...	Page Moderated	NaN	NaN	NaN	Title: 10for10\nPitch: We want you to donate $10 for the 10% of babies by doing 10 days of 10 ex...	Premature birth or significant ongoing medical issues	P07.9 (Prematurity, unspecified)	Chapter P00-P96 - Certain conditions originating in the perinatal period	Not specified	other	Newborn	0-14	Not specified in text; guessing from image: Likely diverse	80.0	85.0	95.0	The text does not specify a particular individual, but rather focuses on supporting families of ...	False	90	90.0	medical expenses, experimental therapies, travel expenses, lost wages	Nationwide	90	85	hopeful	symbolic	False	not_detectable	high	46.500000
8092	8097	Our Harness Hero Princess	Zoe is going for Surgey on the 10th of december where she will be placed into a Spica Cast. Due ...	https://static.givealittle.co.nz/assets/hero/7a00f120-baa5-4374-91be-a9a800a80665-320:False	https://givealittle.co.nz/cause/our-harness-hero-princess	700.0	400.0	closed	Read more	https://givealittle.co.nz/cause/our-harness-hero-princess	30 Nov 2018	1 Dec 2018	Canterbury	7.0	Kia ora whanau and friends!\nA lot of you know that our precious girl, Zoe, has been through a b...	The funds will go directly towards her carseat :)	Created by Abbey Thorne, paying to a verified bank account of_x000D_\nAbbey Thorne (Chambers)	Page Moderated	1.0	Thank you!!\n30 November 2018\nWe are completely overwhelmed by the response we have recieved to...	NaN	Title: Our Harness Hero Princess\nPitch: Zoe is going for Surgey on the 10th of december where s...	Bilateral Developmental Dislocation of the Hips	M47.0	Chapter M00-M99 - Diseases of the musculoskeletal system and connective tissue	Zoe	female	4 months	0-14	Caucasian	90.0	70.0	95.0	The text provides detailed medical history and context, making it highly informative. The urgenc...	True	95	90.0	medical expenses	Canterbury	90	85	hopeful	portrait	True	smiling	high	175.000000
10808	10816	Help Effie	Raising funds for Effie's open heart surgery	https://static.givealittle.co.nz/assets/hero/35a23088-2eb6-41ba-99cc-a67800ed1dc3-320:False	https://givealittle.co.nz/cause/effie	237.0	1000.0	closed	Read more	https://givealittle.co.nz/cause/effie	5 Sep 2016	8 Nov 2016	Auckland	9.0	Little Effie was born needing open heart surgery and without it she faces a very uncertain futur...	NaN	Created by, and paying to a verified bank account of,_x000D_\n _x000D_\n Peter Nicol...	Page Moderated	NaN	NaN	NaN	Title: Help Effie\nPitch: Raising funds for Effie's open heart surgery\nDescription: Little Effi...	Congenital Heart Defect	Q20	Chapter Q00-Q99 - Congenital malformations, deformations and chromosomal abnormalities	Effie	other	indeterminate/unknown	0-14	Caucasian	90.0	70.0	95.0	Age is indeterminate but assumed to be in the 0-14 range due to mention of being a baby. Ethnici...	True	95	90.0	medical expenses, travel expenses, accommodation	Auckland	80	75	hopeful	portrait	True	smiling	high	23.700000

	test	stat	p_value
image_quality	chi2	40.932891	0.0
emotional_tone	chi2	73.290279	0.0
age_group	chi2	21.177083	0.000097
facial_expression	chi2	26.442983	0.000184
ethnicity	chi2	48.378961	0.000376
end	chi2	2186.22469	0.215375
region	chi2	104.213501	0.290067
notes	chi2	2766.80732	0.292727
start	chi2	2427.549619	0.326723
whos_involved	chi2	4433.025955	0.417108
face_visible	chi2	0.606062	0.436275
pitch	chi2	4490.804466	0.489598
title	chi2	4473.047477	0.49277
text	chi2	4499.0	0.49299
actionUri	chi2	4499.0	0.49299
uri	chi2	4499.0	0.49299
hero	chi2	4499.0	0.49299
Charity	chi2	0.407019	0.523486
gender	chi2	1.054881	0.590114
image_type	chi2	7.374276	0.689701
moderated	chi2	0.0	1.0
action	chi2	0.0	1.0
timeLeft	chi2	0.0	1.0

Dep. Variable:	log_amountRaised	No. Observations:	1550
Model:	GLM	Df Residuals:	1541
Model Family:	Gamma	Df Model:	8
Link Function:	Log	Scale:	0.055701
Method:	IRLS	Log-Likelihood:	inf
Date:	Tue, 22 Jul 2025	Deviance:	239.92
Time:	14:45:09	Pearson chi2:	85.8
No. Iterations:	12	Pseudo R-squ. (CS):	nan
Covariance Type:	nonrobust

	coef	std err	z	P>\|z\|	[0.025	0.975]
Intercept	-0.2256	0.167	-1.354	0.176	-0.552	0.101
smiling[T.True]	0.0688	0.013	5.362	0.000	0.044	0.094
gender[T.male]	0.0359	0.013	2.858	0.004	0.011	0.060
gender[T.other]	-0.0900	0.029	-3.095	0.002	-0.147	-0.033
days	0.0002	3.91e-05	5.521	0.000	0.000	0.000
sentiment	0.0035	0.001	6.577	0.000	0.002	0.005
truth	0.0120	0.002	5.936	0.000	0.008	0.016
urgency	0.0097	0.001	12.200	0.000	0.008	0.011
parsed_age	-0.0002	0.000	-0.629	0.529	-0.001	0.000

Dep. Variable:	Target_completion	R-squared:	0.003
Model:	OLS	Adj. R-squared:	-0.002
Method:	Least Squares	F-statistic:	0.6002
Date:	Tue, 22 Jul 2025	Prob (F-statistic):	0.778
Time:	14:45:09	Log-Likelihood:	-13128.
No. Observations:	1550	AIC:	2.627e+04
Df Residuals:	1541	BIC:	2.632e+04
Df Model:	8
Covariance Type:	nonrobust

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	-943.8363	816.545	-1.156	0.248	-2545.493	657.820
smiling[T.True]	51.4457	62.877	0.818	0.413	-71.888	174.780
gender[T.male]	96.6374	61.509	1.571	0.116	-24.012	217.287
gender[T.other]	19.1091	142.511	0.134	0.893	-260.427	298.645
days	-0.0727	0.192	-0.379	0.705	-0.449	0.303
sentiment	0.4612	2.599	0.177	0.859	-4.636	5.558
truth	8.8228	9.898	0.891	0.373	-10.592	28.237
urgency	1.2515	3.907	0.320	0.749	-6.412	8.915
parsed_age	0.3826	1.567	0.244	0.807	-2.690	3.455

Omnibus:	4408.166	Durbin-Watson:	2.007
Prob(Omnibus):	0.000	Jarque-Bera (JB):	117341667.316
Skew:	35.986	Prob(JB):	0.00
Kurtosis:	1349.003	Cond. No.	6.70e+03

Omnibus:	59.312	Durbin-Watson:	1.962
Prob(Omnibus):	0.000	Jarque-Bera (JB):	66.951
Skew:	-0.460	Prob(JB):	2.90e-15
Kurtosis:	3.436	Cond. No.	6.70e+03

	n_estimators	100
	criterion	'squared_error'
	max_depth	None
	min_samples_split	2
	min_samples_leaf	1
	min_weight_fraction_leaf	0.0
	max_features	1.0
	max_leaf_nodes	None
	min_impurity_decrease	0.0
	bootstrap	True
	oob_score	False
	n_jobs	None
	random_state	None
	verbose	0
	warm_start	False
	ccp_alpha	0.0
	max_samples	None
	monotonic_cst	None

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	-7.7171	1.232	-6.264	0.000	-10.134	-5.300
smiling[T.True]	0.5030	0.095	5.302	0.000	0.317	0.689
gender[T.male]	0.2312	0.093	2.491	0.013	0.049	0.413
gender[T.other]	-0.6517	0.215	-3.031	0.002	-1.074	-0.230
days	0.0017	0.000	5.995	0.000	0.001	0.002
sentiment	0.0255	0.004	6.507	0.000	0.018	0.033
truth	0.0735	0.015	4.923	0.000	0.044	0.103
urgency	0.0730	0.006	12.390	0.000	0.061	0.085
parsed_age	-0.0010	0.002	-0.404	0.686	-0.006	0.004