%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)

time: 1.61 s (started: 2025-06-27 20:34:12 +12:00)

df = pd.read_excel("LLM_results_cleaned.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: 3.73 s (started: 2025-06-27 20:34:14 +12:00)

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

gender
female    2258
male      1613
other      540
Name: count, dtype: int64

time: 4.59 ms (started: 2025-06-27 20:34:18 +12:00)

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

time: 1.51 ms (started: 2025-06-27 20:34:18 +12:00)

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

gender
female    2258
male      1613
other      540
Name: count, dtype: int64

time: 7.61 ms (started: 2025-06-27 20:34:18 +12:00)

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

ethnicity
Caucasian                  2541
other                       396
Māori                       272
South Asian                 214
Asian                       161
Pacific Islander            144
Not specified               121
Unknown                     121
African                      57
Filipino                     54
Indian                       49
East Asian                   42
White                        39
Mixed                        38
South Asian/Indian           35
Filipino/Asian               33
Latin American/Hispanic      29
Fijian                       25
Hispanic/Latino              22
Not applicable               18
Name: count, dtype: int64

time: 4.3 ms (started: 2025-06-27 20:34:18 +12:00)

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

ethnicity
Caucasian                  2541
other                       396
Māori                       272
South Asian                 214
Asian                       161
Pacific Islander            144
Not specified               121
Unknown                     121
African                      57
Filipino                     54
Indian                       49
East Asian                   42
White                        39
Mixed                        38
South Asian/Indian           35
Filipino/Asian               33
Latin American/Hispanic      29
Fijian                       25
Hispanic/Latino              22
Not applicable               18
Name: count, dtype: int64

time: 11.5 ms (started: 2025-06-27 20:34:18 +12:00)

df.ethnicity = df.ethnicity.replace({
    "European/Caucasian": "Caucasian",
    "European/White": "Caucasian",
    "European/White": "Caucasian",
    "European/New Zealand": "Caucasian",
    "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                  2541
other                       396
Māori                       272
South Asian                 214
Asian                       161
Pacific Islander            144
Not specified               121
Unknown                     121
African                      57
Filipino                     54
Indian                       49
East Asian                   42
White                        39
Mixed                        38
South Asian/Indian           35
Filipino/Asian               33
Latin American/Hispanic      29
Fijian                       25
Hispanic/Latino              22
Not applicable               18
Name: count, dtype: int64

time: 8.74 ms (started: 2025-06-27 20:34:18 +12:00)

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

condition
cancer                     192
gender dysphoria           135
breast cancer              135
stroke                      75
multiple sclerosis          65
endometriosis               45
bowel cancer                44
type 1 diabetes             39
melanoma                    38
obesity                     37
none specified              36
cerebral palsy              35
terminal cancer             33
infertility                 32
mental health issues        31
covid-19                    27
lung cancer                 27
coronary artery disease     27
hearing loss                26
brain tumor                 26
Name: count, dtype: int64

time: 8.09 ms (started: 2025-06-27 20:34:18 +12:00)

df.sentiment.value_counts()

sentiment
70.0     1717
80.0      603
50.0      524
60.0      493
90.0      422
30.0      258
85.0      235
20.0       64
95.0       35
40.0       28
75.0       26
10.0        5
100.0       1
Name: count, dtype: int64

time: 5.93 ms (started: 2025-06-27 20:34:18 +12:00)

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

use
medical expenses                     3287
travel expenses                       678
experimental therapies                177
lost wages                            141
funeral expenses                       19
funeral costs                          10
medical expenses, travel expenses       5
other                                   4
educational tools                       3
educational expenses                    3
travel expenses, lost wages             3
living costs                            2
legal expenses                          2
financial expenses                      2
living expenses                         2
operational costs                       2
household expenses                      2
educational campaigns                   2
celebration                             1
renovation supplies                     1
Name: count, dtype: int64

time: 6.93 ms (started: 2025-06-27 20:34:18 +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
15-64                    2809
0-14                      649
65+                       558
indeterminate/unknown     395
Name: count, dtype: int64

time: 17.1 ms (started: 2025-06-27 20:34:18 +12:00)

df.gender.value_counts()

gender
female    2258
male      1613
other      540
Name: count, dtype: int64

time: 5.2 ms (started: 2025-06-27 20:34:18 +12:00)

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

time: 2.96 ms (started: 2025-06-27 20:34:18 +12:00)

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

time: 5.88 s (started: 2025-06-27 20:34:18 +12:00)

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

time: 1.95 ms (started: 2025-06-27 20:34:24 +12:00)

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

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

parsed_age
35.0    371
30.0    247
25.0    189
45.0    186
60.0    136
50.0    128
55.0    117
65.0    111
40.0     95
28.0     79
20.0     72
38.0     61
23.0     45
17.0     42
18.0     41
8.0      40
22.0     39
13.0     39
10.0     37
70.0     36
Name: count, dtype: int64

time: 10.8 ms (started: 2025-06-27 20:34:24 +12:00)

df.met_target.value_counts()

met_target
0    3785
1     626
Name: count, dtype: int64

time: 4.99 ms (started: 2025-06-27 20:34:24 +12:00)

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

count    4410.000000
mean       91.346939
std         4.647941
min        50.000000
25%        90.000000
50%        90.000000
75%        95.000000
max       100.000000
Name: deservingness, dtype: float64

time: 11.8 ms (started: 2025-06-27 20:34:24 +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: 644 ms (started: 2025-06-27 20:34:24 +12:00)

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

time: 729 μs (started: 2025-06-27 20:34:25 +12:00)

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

'AIC (Logit): 2447.23'

'AIC (Probit): 2447.64'

time: 106 ms (started: 2025-06-27 20:34:25 +12:00)

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

time: 46.2 ms (started: 2025-06-27 20:34:25 +12:00)

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

time: 45.8 ms (started: 2025-06-27 20:34:25 +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: 512 ms (started: 2025-06-27 20:34:25 +12:00)

list(zip(cols, model.feature_importances_))

[('sentiment', np.float64(0.21365808843272313)),
 ('smiling', np.float64(0.10368107301614912)),
 ('gender', np.float64(0.10246242527274708)),
 ('truth', np.float64(0.2056862734817596)),
 ('urgency', np.float64(0.0559171758768012)),
 ('ethnicity', np.float64(0.05393489406535803))]

time: 11.9 ms (started: 2025-06-27 20:34:25 +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	Target_completion	Charity
0	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	Helping a person in need Helping a person in need Helping out a person in need that would drop h...	Muscle tear leading to internal bleeding and infection	S86.0	Injury, poisoning and certain other consequences of external causes	Tok	male	35	15-64	Pacific Islander	90.0	70.0	95.0	Age estimated based on appearance; ethnicity inferred from image; sentiment reflects a mix of gr...	1	95	60	medical expenses	85.000000	False
1	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	Please help Walter 🙏 to get the veterinary care he needs. Please help Walter 🙏 to get the veteri...	Pancreatitis	K85	11	Walter	male	15	15-64	Unknown	90.0	70.0	95.0	The primary subject is an animal named Walter, not a human. The text mentions pancreatitis as th...	1	90	80	medical expenses	76.000000	False
2	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	Medical Treatment for Atish Lal for Coronary Artery Disease Medical Treatment for Atish Lal for ...	Coronary Artery Disease	I25	I - Diseases of the circulatory system	Atish Lal	male	35	15-64	Indian	90.0	70.0	95.0	Age estimated based on typical age range for coronary artery disease; ethnicity inferred from th...	0	85	50	medical expenses	2.550336	False
3	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	Help me to reclaim my freedom. Help me to reclaim my freedom. Help my to reclaim my freedom and ...	Long Covid, Chronic Fatigue/ME, PEM, Acquired Apraxia of Speech	G93.3 (Postviral fatigue syndrome)	Chapter VI: Diseases of the nervous system	Amy	female	35	15-64	Caucasian	80.0	60.0	90.0	Age estimated based on appearance in the image. Ethnicity assumed based on visual characteristics.	1	90	70	medical expenses	49.983333	False
4	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	Help My Mum Fight Cancer - Every Bit Counts Help My Mum Fight Cancer - Every Bit Counts Rasing f...	Cancer	C00-C97	Neoplasms	Not explicitly mentioned	female	Not explicitly mentioned	indeterminate/unknown	Indian	90.0	80.0	95.0	Age and name not explicitly provided in the text. Ethnicity assumed based on the mention of trea...	0	95	50	medical expenses	1.260000	False
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
4406	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	New Zealand MindBody Trust web-based 'whole person' resources for healthworkers and their patien...	Chronic Disorders	R53	Chapter XVIII: Symptoms, Signs and Abnormal Clinical and Laboratory Findings, Not Elsewhere Clas...	Not specified	other	Not specified	indeterminate/unknown	Not specified	80.0	90.0	95.0	The text does not specify a particular individual, so details like name, gender, age, and ethnic...	0	90	50	medical expenses	74.444444	True
4407	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	Diane's cyst removal to continue my dream job Diane's cyst removal to continue my dream job Larg...	Radial Ganglion Cyst	M75.3	M00-M99 Diseases of the musculoskeletal system and connective tissue	Diane	female	45	15-64	Caucasian	80.0	50.0	90.0	Age estimated based on appearance; ethnicity inferred from image.	0	85	60	medical expenses	14.600000	False
4408	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	A thank you to Waikato NICU for helping so many little babies and their families A thank you to ...	Premature birth	P07	Chapter XVII: Congenital malformations, deformations and chromosomal abnormalities	Charlie and Oliver	male	Newborn (preterm)	0-14	Asian	90.0	80.0	95.0	The text mentions premature birth and the need for a portable cardiac monitor for preterm infant...	1	95	70	medical expenses	25.464857	False
4409	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	Help restore David’s smile after being assaulted Help restore David’s smile after being assaulte...	Dental Trauma	S02.5	Injury, poisoning and certain other consequences of external causes	David Shanhun	male	Not specified	15-64	Caucasian	80.0	70.0	90.0	Age is not explicitly mentioned in the text; assumed to be within the 15-64 range based on conte...	0	90	70	medical expenses	14.766972	False
4410	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	TRUS Biopsy Machine for Taranaki TRUS Biopsy Machine for Taranaki The men of Taranaki need this ...	Prostate Cancer	C61	Neoplasms	Men of Taranaki	male	Not specified	15-64	other	80.0	90.0	95.0	The text does not specify a specific individual but refers to 'men of Taranaki'. Age is inferred...	0	90	50	medical expenses	9.550000	True

	test	stat	p_value
age_group	chi2	36.113648	0.0
ethnicity	chi2	46.450561	0.000428
end	chi2	2164.655553	0.254419
gender	chi2	2.247197	0.325108
start	chi2	2401.026199	0.382451
whos_involved	chi2	4353.107405	0.411837
title	chi2	4391.565606	0.469235
pitch	chi2	4402.788284	0.489563
text	chi2	4411.0	0.492921
hero	chi2	4411.0	0.492921
uri	chi2	4411.0	0.492921
actionUri	chi2	4411.0	0.492921
notes	chi2	2968.234961	0.607771
Charity	chi2	0.170503	0.679666
action	chi2	0.0	1.0
timeLeft	chi2	0.0	1.0
moderated	chi2	0.0	1.0

Dep. Variable:	['met_target[False]', 'met_target[True]']	No. Observations:	3144
Model:	GLM	Df Residuals:	3136
Model Family:	Binomial	Df Model:	7
Link Function:	Logit	Scale:	1.0000
Method:	IRLS	Log-Likelihood:	-1215.6
Date:	Fri, 27 Jun 2025	Deviance:	2431.2
Time:	20:34:25	Pearson chi2:	3.15e+03
No. Iterations:	5	Pseudo R-squ. (CS):	0.03019
Covariance Type:	nonrobust

	coef	std err	z	P>\|z\|	[0.025	0.975]
Intercept	5.7990	1.485	3.906	0.000	2.889	8.709
smiling[T.True]	-0.3097	0.115	-2.694	0.007	-0.535	-0.084
gender[T.male]	-0.1971	0.110	-1.794	0.073	-0.412	0.018
gender[T.other]	0.7005	0.315	2.227	0.026	0.084	1.317
sentiment	-0.0203	0.005	-4.357	0.000	-0.029	-0.011
truth	-0.0359	0.018	-2.010	0.044	-0.071	-0.001
urgency	0.0060	0.007	0.910	0.363	-0.007	0.019
parsed_age	0.0139	0.003	4.522	0.000	0.008	0.020

Dep. Variable:	['met_target[False]', 'met_target[True]']	No. Observations:	3144
Model:	GLM	Df Residuals:	3136
Model Family:	Binomial	Df Model:	7
Link Function:	Probit	Scale:	1.0000
Method:	IRLS	Log-Likelihood:	-1215.8
Date:	Fri, 27 Jun 2025	Deviance:	2431.6
Time:	20:34:25	Pearson chi2:	3.16e+03
No. Iterations:	6	Pseudo R-squ. (CS):	0.03006
Covariance Type:	nonrobust

Dep. Variable:	Target_completion	R-squared:	0.001
Model:	OLS	Adj. R-squared:	-0.001
Method:	Least Squares	F-statistic:	0.5881
Date:	Fri, 27 Jun 2025	Prob (F-statistic):	0.766
Time:	20:34:25	Log-Likelihood:	-25776.
No. Observations:	3144	AIC:	5.157e+04
Df Residuals:	3136	BIC:	5.162e+04
Df Model:	7
Covariance Type:	nonrobust

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	-136.3356	387.947	-0.351	0.725	-896.991	624.320
smiling[T.True]	17.3750	33.570	0.518	0.605	-48.446	83.196
gender[T.male]	53.2272	33.259	1.600	0.110	-11.985	118.440
gender[T.other]	-25.6309	80.636	-0.318	0.751	-183.736	132.474
sentiment	0.6093	1.214	0.502	0.616	-1.771	2.990
truth	0.3478	4.473	0.078	0.938	-8.422	9.117
urgency	1.4226	2.067	0.688	0.491	-2.629	5.474
parsed_age	-0.5585	0.901	-0.620	0.536	-2.325	1.209

Omnibus:	9405.879	Durbin-Watson:	2.003
Prob(Omnibus):	0.000	Jarque-Bera (JB):	530902384.962
Skew:	42.477	Prob(JB):	0.00
Kurtosis:	2014.336	Cond. No.	3.63e+03

Dep. Variable:	met_target	R-squared:	0.029
Model:	OLS	Adj. R-squared:	0.026
Method:	Least Squares	F-statistic:	13.19
Date:	Fri, 27 Jun 2025	Prob (F-statistic):	7.41e-17
Time:	20:34:25	Log-Likelihood:	-1072.2
No. Observations:	3144	AIC:	2160.
Df Residuals:	3136	BIC:	2209.
Df Model:	7
Covariance Type:	nonrobust

Omnibus:	1102.280	Durbin-Watson:	2.006
Prob(Omnibus):	0.000	Jarque-Bera (JB):	2752.352
Skew:	2.002	Prob(JB):	0.00
Kurtosis:	5.230	Cond. No.	3.63e+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	z	P>\|z\|	[0.025	0.975]
Intercept	3.0633	0.771	3.971	0.000	1.551	4.575
smiling[T.True]	-0.1667	0.061	-2.711	0.007	-0.287	-0.046
gender[T.male]	-0.1131	0.060	-1.890	0.059	-0.230	0.004
gender[T.other]	0.3691	0.161	2.297	0.022	0.054	0.684
sentiment	-0.0104	0.002	-4.293	0.000	-0.015	-0.006
truth	-0.0187	0.009	-2.032	0.042	-0.037	-0.001
urgency	0.0038	0.004	1.045	0.296	-0.003	0.011
parsed_age	0.0078	0.002	4.687	0.000	0.005	0.011

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	-0.2041	0.150	-1.360	0.174	-0.498	0.090
smiling[T.True]	0.0358	0.013	2.760	0.006	0.010	0.061
gender[T.male]	0.0241	0.013	1.872	0.061	-0.001	0.049
gender[T.other]	-0.0751	0.031	-2.409	0.016	-0.136	-0.014
sentiment	0.0020	0.000	4.322	0.000	0.001	0.003
truth	0.0033	0.002	1.919	0.055	-7.21e-05	0.007
urgency	-0.0008	0.001	-1.001	0.317	-0.002	0.001
parsed_age	-0.0016	0.000	-4.564	0.000	-0.002	-0.001