Dataset Viewer
Auto-converted to Parquet Duplicate
step
int64
type
string
amount
float64
nameOrig
string
oldbalanceOrg
float64
newbalanceOrig
float64
nameDest
string
oldbalanceDest
float64
newbalanceDest
float64
isFraud
int64
isFlaggedFraud
int64
278
CASH_IN
330,218.42
C632336343
20,866
351,084.42
C834976624
452,419.57
122,201.15
0
0
15
PAYMENT
11,647.08
C1264712553
30,370
18,722.92
M215391829
0
0
0
0
10
CASH_IN
152,264.21
C1746846248
106,589
258,853.21
C1607284477
201,303.01
49,038.8
0
0
403
TRANSFER
1,551,760.63
C333676753
0
0
C1564353608
3,198,359.45
4,750,120.08
0
0
206
CASH_IN
78,172.3
C813403091
2,921,331.58
2,999,503.88
C1091768874
415,821.9
337,649.6
0
0
259
PAYMENT
915.13
C2002954533
0
0
M290849763
0
0
0
0
188
CASH_OUT
20,603.87
C813757373
0
0
C823291717
558,068.66
578,672.53
0
0
139
CASH_OUT
58,605.72
C1850864812
0
0
C618657299
585,494.94
644,100.66
0
0
230
PAYMENT
4,865.11
C886849972
0
0
M623175144
0
0
0
0
544
CASH_OUT
118,131.63
C390714641
0
0
C366360355
8,131,691.35
8,476,246.86
0
0
45
CASH_OUT
141,100.88
C1514989792
80,506
0
C409578677
89,384.09
230,484.96
0
0
163
CASH_OUT
384,177.48
C94444270
40,348
0
C379896225
0
384,177.48
0
0
211
CASH_OUT
153,486.36
C513018655
20,476
0
C1302281681
85,173.16
238,659.53
0
0
302
CASH_OUT
51,042.59
C122114408
184
0
C705120650
380,143.14
431,185.74
0
0
18
PAYMENT
48,691.16
C1810046508
0
0
M487081598
0
0
0
0
322
PAYMENT
36,598.65
C1365498607
8,240.66
0
M969218364
0
0
0
0
158
CASH_OUT
124,443.29
C12524561
14,236
0
C579157353
0
124,443.29
0
0
131
CASH_OUT
510,190.97
C1564771551
57,590
0
C718731643
115,914.88
626,105.86
0
0
253
TRANSFER
940,927.27
C1346570604
8,568
0
C1178934515
0
940,927.27
0
0
132
CASH_OUT
240,156.34
C807101143
0
0
C1531623971
2,811,902.28
3,052,058.62
0
0
283
CASH_OUT
44,360.26
C346953986
0
0
C526046232
282,664.52
327,024.78
0
0
43
CASH_OUT
338,458.24
C25552315
30,283
0
C75671301
1,055,729.15
1,564,103.49
0
0
642
CASH_IN
311,748.24
C200195468
4,279
316,027.24
C786997986
27,843.39
0
0
0
164
DEBIT
151.58
C1576094043
10,659
10,507.42
C1308477441
50,706.33
50,857.9
0
0
326
PAYMENT
26,565.95
C1512675672
0
0
M1330826048
0
0
0
0
132
CASH_OUT
6.57
C1286124147
0
0
C1885613878
2,909,681.35
2,909,687.93
0
0
37
CASH_IN
23,898.77
C1953063740
50,271
74,169.77
C212080696
205,872.37
181,973.6
0
0
9
CASH_OUT
26,360.78
C470767940
365
0
C598132143
202,458.96
244,056.68
0
0
157
CASH_IN
192,935.2
C709178804
51,937
244,872.2
C1163988187
171,874.85
0
0
0
406
TRANSFER
18,801.43
C920671013
126
0
C1418884586
0
18,801.43
0
0
235
PAYMENT
3,516.08
C633099116
0
0
M2117267163
0
0
0
0
346
PAYMENT
6,987.8
C220437223
0
0
M1751641542
0
0
0
0
19
CASH_IN
8,808.73
C1717919401
50,485
59,293.73
C645140627
177,448.43
168,639.7
0
0
208
PAYMENT
24,649.07
C1353964860
0
0
M1059671726
0
0
0
0
332
CASH_OUT
97,562.73
C1894828300
5,396
0
C1975390043
2,455,585.33
2,553,148.05
0
0
139
TRANSFER
598,680.17
C544409289
45,059
0
C956338890
2,559,233
3,157,913.17
0
0
304
CASH_OUT
218,659.82
C88319926
1,005
0
C1971781194
16,320,325.66
16,592,681.56
0
0
686
CASH_IN
304,408.97
C893917810
4,276,845.77
4,581,254.74
C1327824794
1,443,186.93
1,314,479.93
0
0
139
PAYMENT
5,090.28
C1405934398
50,617
45,526.72
M839389189
0
0
0
0
226
CASH_OUT
176,016.04
C800760258
0
0
C358043414
12,916,654.7
13,005,481.5
0
0
201
PAYMENT
11,605.7
C484951468
0
0
M1504624467
0
0
0
0
14
CASH_IN
234,850.43
C982073663
669,990.79
904,841.22
C1216254870
298,639.42
49,936.17
0
0
323
CASH_OUT
186,549.79
C1647129298
0
0
C1714837024
242,473.16
429,022.95
0
0
225
CASH_OUT
79,042.15
C2086229320
6,064
0
C1654662784
410,360.71
489,402.86
0
0
138
PAYMENT
5,055.72
C315487150
108,809.83
103,754.11
M1717055694
0
0
0
0
41
PAYMENT
12,020.77
C1235787929
639,020.83
627,000.06
M1043686153
0
0
0
0
399
CASH_IN
123,326.84
C979213321
5,477,362
5,600,688.84
C787242927
3,470,716.16
3,347,389.32
0
0
120
PAYMENT
14,241.21
C1653380908
109,207
94,965.79
M77882578
0
0
0
0
373
PAYMENT
23,480.16
C1099535493
5,309
0
M933507080
0
0
0
0
179
CASH_IN
195,205.19
C2147063789
1,961,856.75
2,157,061.94
C1021587874
1,150,566.93
955,361.74
0
0
544
CASH_OUT
145,319.54
C237037918
8,056
0
C95318372
696,100.99
841,420.53
0
0
369
PAYMENT
14,270.03
C1496903190
3,309
0
M946866772
0
0
0
0
275
PAYMENT
6,851.17
C1051106607
0
0
M1348761534
0
0
0
0
396
CASH_IN
296,959.67
C731717743
10,609
307,568.67
C148906605
0
0
0
0
304
CASH_IN
166,630.9
C1410677661
10,944
177,574.9
C704312950
240,525
73,894.11
0
0
162
CASH_IN
165,135.41
C677811242
187,326.54
352,461.95
C788874634
190,273.97
25,138.56
0
0
304
PAYMENT
1,704.03
C74556831
0
0
M1951402325
0
0
0
0
22
CASH_IN
390,903.09
C41573174
7,947,520.93
8,338,424.03
C860929789
707,650.38
316,747.29
0
0
229
PAYMENT
21,061.24
C40966245
0
0
M1653653617
0
0
0
0
259
PAYMENT
8,129.46
C40866646
0
0
M182944108
0
0
0
0
234
PAYMENT
6,492.32
C1321671140
252,978.2
246,485.88
M533921438
0
0
0
0
304
CASH_OUT
16,242.22
C1322294401
0
0
C160609026
1,401,840.75
1,418,082.97
0
0
352
CASH_OUT
286,864.49
C998029122
0
0
C1563899518
598,674.63
885,539.12
0
0
139
CASH_IN
147,351.57
C710908433
6,185,044.14
6,332,395.71
C508488713
7,292,368.53
7,145,016.96
0
0
229
TRANSFER
21,362.19
C1977482177
0
0
C1554823726
102,800.19
124,162.37
0
0
186
PAYMENT
16,871.75
C1345262134
374,693.53
357,821.78
M1183610480
0
0
0
0
406
CASH_OUT
88,826.43
C952906355
249,733
160,906.57
C1084012188
963,113.59
1,051,940.02
0
0
469
PAYMENT
38,509.2
C174212239
0
0
M1265795703
0
0
0
0
399
PAYMENT
7,301.4
C239144018
21,426
14,124.6
M1684201690
0
0
0
0
288
PAYMENT
51,907.16
C1577059714
48,529
0
M1955312313
0
0
0
0
204
PAYMENT
343.06
C1370498771
0
0
M301320628
0
0
0
0
232
CASH_OUT
24,413.46
C1542135254
5,088
0
C1938475526
0
24,413.46
0
0
192
CASH_OUT
125,568.72
C923977418
55,360
0
C779506402
0
125,568.72
0
0
13
CASH_OUT
203,321.66
C1009192782
7,707.02
0
C1015529677
223,460.41
426,782.08
0
0
281
PAYMENT
7,452.27
C1418729422
254,793
247,340.73
M1996394103
0
0
0
0
379
CASH_IN
117,253.94
C2029840344
1,527,903.53
1,645,157.47
C2142547587
309,923.06
192,669.12
0
0
159
CASH_OUT
16,500.28
C731563967
0
0
C914957636
1,039,511.46
1,056,011.74
0
0
324
PAYMENT
25,380.11
C1515172243
0
0
M849428673
0
0
0
0
230
CASH_OUT
95,623.45
C1025174837
95,785
161.55
C1337597412
0
95,623.45
0
0
405
CASH_IN
122,899.15
C1358412469
6,017
128,916.15
C1995332832
227,088.65
104,189.5
0
0
380
CASH_IN
255,661.72
C1144246806
1,148,679.19
1,404,340.91
C87426661
1,991,419.36
1,735,757.64
0
0
307
CASH_OUT
124,210.26
C583984242
13,648
0
C408688303
637,289.41
761,499.67
0
0
43
TRANSFER
406,390.63
C449334349
0
0
C1477190619
596,285.7
1,002,676.33
0
0
404
CASH_IN
12,747.45
C1946500289
3,027,551.36
3,040,298.81
C748382555
193,795.69
181,048.24
0
0
36
PAYMENT
17,210.24
C2035379457
468,747.37
451,537.13
M1250735899
0
0
0
0
401
CASH_IN
287,548.26
C2147246002
6,300,194.15
6,587,742.41
C1034690288
1,211,290.25
923,741.99
0
0
259
CASH_IN
326,712.9
C1906310517
10,763,635.06
11,090,347.96
C809795627
2,415,359.64
2,088,646.74
0
0
228
PAYMENT
241.89
C1249789290
103,317.6
103,075.72
M870078784
0
0
0
0
236
PAYMENT
5,883.37
C1938448095
0
0
M1022031461
0
0
0
0
297
PAYMENT
13,055.28
C508385511
0
0
M956027336
0
0
0
0
12
PAYMENT
616.88
C1013341637
107,243.5
106,626.62
M1942028884
0
0
0
0
236
CASH_OUT
107,759.5
C1408936980
0
0
C1403087041
1,507,711.83
1,615,471.33
0
0
253
CASH_IN
18,868.29
C1363041822
21,137,221.01
21,156,089.3
C1761313958
34,831.61
15,963.32
0
0
13
PAYMENT
4,986.63
C1359889730
0
0
M1124177815
0
0
0
0
36
PAYMENT
17,692.64
C1270549119
10,268
0
M743266449
0
0
0
0
403
PAYMENT
7,457.89
C2051931889
0
0
M2079068430
0
0
0
0
300
DEBIT
2,262.52
C841098437
49,271
47,008.48
C1097907228
124,780
127,042.52
0
0
564
CASH_IN
67,418.91
C1484852473
52,181
119,599.91
C170358894
1,659,195.46
1,591,776.55
0
0
254
PAYMENT
9,414.76
C2023654274
5,839.9
0
M210283336
0
0
0
0
206
PAYMENT
17,485.51
C928494081
9,339
0
M981647331
0
0
0
0
End of preview. Expand in Data Studio

PaySim Financial Fraud Detection — EDA & Dataset Analysis

Presentation Video

Dataset Overview

This project analyzes the PaySim Financial Fraud Detection dataset from Kaggle (source: chitwanmanchanda/fraudulent-transactions-data). The dataset contains simulated mobile money transactions with fraud labels, representing 100,000 randomly sampled transactions (random_state=42) from the original 6,362,620 rows.

Key characteristics:

  • Total records: 100,000 transactions
  • Target variable: isFraud (0 = legitimate, 1 = fraudulent)
  • Data quality issues found & resolved: 200 missing amount values, 300 missing oldbalanceOrg values, 500 duplicate rows, 400 inconsistent type formatting
  • Class imbalance: Fraud is extremely rare (~0.14% of all transactions)

Research Question

"How do specific economic indicators, such as transaction types and sudden inconsistencies between account balances, act as financial fingerprints to accurately predict fraudulent activity, and can identifying these logical gaps in the data lead to more efficient and cost-effective detection systems?"

This project investigates whether patterns in transaction characteristics—particularly transaction type, amount magnitude, and account balance behavior—can serve as reliable fraud indicators without relying on demographic or user-level features.


Features

Feature Type Description
step int Time step of the transaction (1-743, representing hours in simulation)
type str Transaction type: CASH_IN, CASH_OUT, DEBIT, PAYMENT, TRANSFER
amount float Transaction amount (in currency units)
oldbalanceOrg float Original account balance before transaction
newbalanceOrig float Original account balance after transaction
oldbalanceDest float Destination account balance before transaction
newbalanceDest float Destination account balance after transaction
isFraud int Binary fraud label (0 = legitimate, 1 = fraudulent)
isFlaggedFraud int Whether transaction was flagged by bank's system
amount_log float Log-transformed amount (log1p) for analysis

EDA Methodology

1. Data Loading & Cleaning

Issues introduced to simulate real-world financial data quality problems:

  • 200 missing values in amount (simulating incomplete transaction records)
  • 300 missing values in oldbalanceOrg (simulating balance data not captured at logging time)
  • 500 duplicate rows (simulating double-logging errors in the payment system)
  • 400 rows with inconsistent casing in type (simulating data entry errors)

Cleaning steps applied:

  • Missing amount → dropped rows: A transaction with no amount is unusable and cannot be safely imputed
  • Missing oldbalanceOrg → filled with median: Median chosen over mean because the amount column is right-skewed; median is a robust estimator
  • Duplicates → drop_duplicates(): Removed all 500 duplicate rows to prevent inflated counts
  • type inconsistency → .str.upper().str.strip(): Standardized all 5 transaction types to uppercase to prevent transfer and TRANSFER being treated as separate categories

2. Outlier Detection & Treatment

Applied Interquartile Range (IQR) method on amount feature. 5,358 outliers detected (5.4% of data). Decision: retained all outliers because fraud transactions often occur at extreme values.

Outlier Detection — Amount Distribution

Applied log1p transformation to amountamount_log for cleaner visualization.

Log-Transformed Amount Distribution

3. Descriptive Statistics

Calculated summary statistics by fraud status. Key insight: fraud transactions are drastically larger than legitimate ones — average fraud amount ~1,340,000 vs. legitimate ~179,000 (ratio: 7.5x larger).

4. Correlation Analysis

Generated correlation heatmap of all numerical features to identify relationships between balance features and fraud status.

Correlation Heatmap

5. Transaction Type Distribution

Visualized how transactions are distributed across all 5 types.

Transaction Type Distribution

6. Fraud by Transaction Type (Raw Count)

Showed where fraud actually occurs across transaction types before normalizing for volume.

Fraud Count by Transaction Type


Key Findings

Overall Fraud Characteristics

  • Fraud rate: 0.14% of all transactions (142 fraudulent out of 100,000)
  • Class imbalance: Highly skewed toward legitimate transactions
  • Amount severity: Fraudulent transactions are 7.5x larger on average
  • Bank's detection: Only 0.56% of fraudulent transactions flagged (isFlaggedFraud=1)

Transaction Type Distribution

  • Most common type: TRANSFER (35%)
  • PAYMENT transactions: 31%
  • CASH_OUT: 24%
  • DEBIT: 8%
  • CASH_IN: 2%

Fraud Distribution by Type

  • TRANSFER: 0.99% fraud rate (77 frauds out of 7,806 transfers)
  • CASH_OUT: 0.30% fraud rate (65 frauds out of 23,235 cash-outs)
  • PAYMENT: 0% fraud rate (0 frauds)
  • DEBIT: 0% fraud rate (0 frauds)
  • CASH_IN: 0% fraud rate (0 frauds)

Four Research Questions & Insights

Research Q1: "Do high-value transactions carry higher fraud risk?"

Finding: Yes, fraud transactions occupy the extreme upper tail of the amount distribution. The highest-value transactions show disproportionately high fraud rates, indicating that transaction magnitude is a critical risk indicator.

Q1 — Amount Distribution by Fraud Status

Conclusion: Amount is a strong univariate fraud signal. Fraudsters tend to target high-value transfers to maximize stolen funds per transaction.


Research Q2: "Do fraudsters drain the sender's account to zero?"

Finding:

  • Fraudsters drain the sender account to zero in ~96% of fraud cases
  • Legitimate transactions show account depletion in only ~0.3% of cases
  • This is a massive logical inconsistency: legitimate users rarely empty accounts, but fraudsters consistently do

Q2 — Account Drain Rate: Fraud vs Legitimate

Conclusion: Balance depletion is one of the strongest fraud indicators in the dataset. This "emptying the well" behavior is a logical fingerprint of fraudulent activity — attackers maximize extraction and disappear.


Research Q3: "Are certain transaction types fraud-free?"

Finding:

  • TRANSFER and CASH_OUT are the only types with any fraud (0.99% and 0.30% respectively)
  • PAYMENT, DEBIT, and CASH_IN have a 0% fraud rate in this dataset
  • Fraud is concentrated in transaction types that move money away from the original account

Q3 — Fraud Rate (%) by Transaction Type

Conclusion: Transaction type is a critical categorical predictor. A simple rule-based filter (flag TRANSFER & CASH_OUT only) would catch 100% of fraud with minimal false positives.


Research Q4: "Is fraud correlated with transaction amount ranges?"

Finding:

  • Small (<1K): ~0.04% fraud rate
  • Medium (1K–10K): ~0.10% fraud rate
  • Large (10K–100K): ~0.27% fraud rate
  • Very Large (>100K): ~1.08% fraud rate
  • Clear monotonic increase: Larger amount buckets have exponentially higher fraud risk

Q4 — Fraud Rate (%) by Transaction Amount Range

Conclusion: Amount binning creates a simple but effective fraud scoring feature. The largest transactions contain ~8x more fraud than the smallest, making amount a strong standalone predictor.


Key Decisions & Rationale

Why keep outliers?

Fraud is inherently an outlier behavior. Removing high-value transactions would eliminate the most suspicious cases. Decision: Retain all 5,358 outliers because they represent the exact transactions we want to catch.

Why log-transform the amount?

Raw amounts span from 0 to millions, creating extreme skewness. Log transformation makes distributions more interpretable, reduces visual dominance of extreme values, and improves visualization clarity without losing information.

Why 100K sample?

The original dataset has 6.3M rows. A 100K random sample maintains the representative fraud rate (~0.14%), reduces computational overhead, and is sufficient for EDA. Reproducibility is ensured with seed=42.

Why focus on these 4 research questions?

Each addresses a different dimension of fraud detection logic — amount risk, balance consistency, transaction-type filtering, and amount-bin stratification. Together they reveal that fraud has multiple independent signals that all point in the same direction.


Conclusion

This EDA reveals that financial fraud leaves multiple consistent fingerprints in transaction data. Rather than relying on complex models or external features, we can identify fraud through transaction type filtering (TRANSFER & CASH_OUT only), amount thresholds (very large transactions are far more likely to be fraud), and balance logic checks (account depletion to zero is virtually diagnostic).

These insights suggest that cost-effective fraud detection is possible through simple rule-based systems with high precision, tiered monitoring for high-risk transaction types, and real-time balance anomaly detection. Future work could focus on ensemble methods combining these signals, time-series analysis, and network-level features to achieve even higher detection rates.


Dataset source: Kaggle — PaySim Financial Fraud Detection
Sample method: Random sampling, n=100,000, random_state=42
Analysis date: April 2026


This project was created for educational purposes only and is submitted as part of a Data Science course assignment at Reichman University.

Downloads last month
106