Overview
This library contains tools for evaluating fidelity and privacy of synthetic data.
Usage
Import the desired modules from the library:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from tonic_reporting import univariate, multivariate, privacy
Preface
Numeric columns refer to columns encoded as numeric. Numerical data types in the schema underlying a model may be encoded as other types.
Categorical columns refer to columns encoded as categorical.
source_df is a Pandas DataFrame of original data from the source database
synth_df is a Pandas DataFrame of sampled data from trained models
The source and synthetic DataFrames should be equal in row count and schema.
Numeric Column Statistics
univariate.summarize_numeric(source_df, synth_df, numeric_cols)
Categorical Column Statistics
univariate.summarize_categorical(source_df, synth_df, categorical_cols)
Numeric Column Comparative Histograms
fig, axarr = plt.subplots(1, len(numeric_cols), figsize = (9,12))
axarr = axarr.ravel()
for col, ax in zip(numeric_cols, axarr):
univariate.plot_histogram(source_df, synth_df, col,ax)
Categorical Column Comparative Frequency Tables
for col in categorical_cols:
univariate.plot_frequency_table(source_df, synth_df, col, ax)
Numeric Column Aggregates Over Time
If the data represents time series, we can visualize means and confidence intervals of numeric features over time:
for col in numeric_cols:
fig, ax = plt.subplots(figsize=(10, 8))
univariate.plot_events_means(source_df, synth_df, col, order_col, ax=ax)
and
for col in numeric_cols:
fig, ax = plt.subplots(figsize=(12, 10))
univariate.plot_events_confidence_intervals(source_df, synth_df, col, order_col, ax=ax)
where order_col denotes the time/order column.
Numeric Column Multivariate Correlations Table
multivariate.summarize_correlations(source_df, synth_df, numeric_cols)
Numeric Column Multivariate Correlations Heat Map
fig, axarr = plt.subplots(1, 2, figsize=(13, 8))
multivariate.plot_correlations(source_df, synth_df, numeric_cols, axarr=axarr, )
fig.tight_layout()
Distance to Closest Record Comparison
syn_dcr, real_dcr = privacy.compute_dcr(source_df, synth_df, numeric_cols, categorical_cols)
fig, ax = plt.subplots(1,1,figsize=(8,6))
ax.hist(real_dcr,bins=300,label = 'Real vs. real', color='mediumpurple');
ax.hist(syn_dcr,bins=300,label='Synthetic vs. real', color='mediumturquoise');
ax.tick_params(axis='both', which='major', labelsize=14)
ax.set_title('Distances to closest record',fontsize=22)
ax.legend(fontsize=16);
Metadata
Release files for tonic-reporting 1.5.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| tonic_reporting-1.5.0.tar.gz | 10.2 kB | Without provenance | Details for tonic_reporting-1.5.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| tonic_reporting-1.5.0-py3-none-any.whl | Python 3 | none | any | Details for tonic_reporting-1.5.0-py3-none-any.whl |
Total release size: 21.9 kB
Release files / tonic_reporting-1.5.0.tar.gz
| Download URL | tonic_reporting-1.5.0.tar.gz |
|---|---|
| Size | 10.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
d4cceba407ddd4fac3ea5a97fb0467cbbe9df8b4bf19f97d15811165cd4b144f
|
|
BLAKE2b-256 checksum How to use checksums |
5dabc3bced69de1019f38201a6fe2a6ce096001f470391e3a717fc51948d3f25
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.10.9 Darwin/21.6.0
|
Release files / tonic_reporting-1.5.0-py3-none-any.whl
| Download URL | tonic_reporting-1.5.0-py3-none-any.whl |
|---|---|
| Size | 11.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
588579a4b761af97a5e7acb811a4a962b2adfbd2ccdeed946a1c0e755105d955
|
|
BLAKE2b-256 checksum How to use checksums |
dad704ad17652efcf0c4e5f2ed82d16ab321703ef447b1c9cf6444912ac20368
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.10.9 Darwin/21.6.0
|
Security for tonic-reporting 1.5.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.