bp is my personal library for day to day use scripts such as data cleaning, machine learning, data visualization and so on.
This module contains various data science and machine learning methods, which can seen in GIF videos in 'examples/gif' page.
The module "bp" expands pandas DataFrame API and adds various visualization and data analysis functionalities. For example, we can get a plot of a numeric column using the method "df.bp.plot_num("my_numeric_variable")".
| BPAccessor | hlp | plot_daily_cat | plotly_corr_heatmap |
| Plotly_Charts | json | plot_date_cat | plotly_countplot |
| RandomColor | light_axis | plot_ecdf | plotly_country_plot |
| add_interactions | lm_plot | plot_gini | plotly_distplot |
| add_text_barplot | lm_residual_corr_plot | plot_ks | plotly_histogram |
| adjustedR2 | lm_stats | plot_multiple_jointplots_with_pearsonr | plotly_mapbox |
| hex_to_rgb | magnify | plot_num | plotly_radar_plot |
| discrete_cmap | multiple_linear_regression | plot_num_cat | plotly_scattergl_plot |
| display_calendar | no_axis | plot_num_cat2 | plotly_scattergl_plot_colorcol |
| plot_plot_binn | optimize_memory | plot_num_num | plotly_scattergl_plot_subplots |
| find_corr | parallelize_dataframe | plot_pareto | plotly_usa_bubble_map |
| freq_count | parse_json_col | plot_roc_auc | plotly_usa_map |
| get_binary_classification_report | partial_corr | plot_roc_skf | plotly_usa_map2 |
| get_binary_classification_scalar_metrics | plot_boxplot_cats_num | plot_simple_linear_regression | point_biserial_correlation |
| get_binary_classification_scalar_metrics2 | plot_cat | plot_statistics | print_calendar |
| get_column_descriptions | plot_cat_binn | plot_stem | print_confusion_matrix |
| get_distinct_colors | plot_cat_cat | plot_two_clusters | print_df_eval |
| get_false_negative_frauds | plot_cat_cat2 | plotly_agg_country_plot | print_statsmodels_summary |
| get_high_correlated_features_df | plot_cat_cat_pct | plotly_agg_usa_plot | random |
| get_mpl_style | plot_cat_num | plotly_binary_clf_evaluation | regression_residual_plots |
| get_outliers | plot_confusion_matrix_plotly | plotly_boxplot | remove_outliers |
| get_outliers_kde | plot_corr | plotly_boxplot_allpoints_with_outliers | rgb2hex |
| get_plotly_colorscale | plot_corr_style | plotly_boxplot_categorical_column | select_kbest_features |
| get_yprobs_sorted_proportions | plot_corrplot_with_pearsonr | plotly_bubbleplot | show_methods |
For the usage of the module, visit API tutorials at ReadTheDocs or Github Pages.
This repository exists for two reasons.
- Build a code base for personal use so that I don't have to write the same code snippet twice.
- Learn the best practices in software industry.
Live Example
We can run the bp Jupyter notebooks live over the web at Binder:
Rendered Preview
| Notebook | Rendered | Description | Author |
|---|---|---|---|
| demo.ipynb | ipynb, rendered | Bhishan Poudel |
- Bhishan Poudel (Ph.D Physics, Ohio University)
Go to your conda environment where you want to install this module, and then install the module.
- First go the folder where is setup.py located.
ls # there must be setup.py file.
which pip # make sure you are in right environment
python setup.py install --user # This will build some eggs and build directories.- Once we install the module, we can not change the source codes.
- This command will install all the modules given in
setup.pyfile.
- If we want to change the source code and update the module, we can install developer version.
which pip # make sure you are in correct conda environment
pip install -e . # -e means editable version and dot is the path
# where there is setup.py file located.
# this will create: bp.egg-info folder.
# make sure you ignore this folder in .gitignore# step 00: Go to the path where you want to work on
mkdir -p ~/temp; cd ~/temp
mkdir example_docker; cd example_docker
# step 01: download the Dockerfile from this repository
wget https://raw.githubusercontent.com/bhishanpdl/bp/master/Dockerfile
clear;ls
cat Dockerfile
# step 02: Run the docker in your mac (you will see docker icon on menubar)
# step 03: Build the docker image and give name "bp"
# this takes about 2 minutes
# this installs vim, jupyter and some latex modules needed for jupyter-notebook
# --no-cache will delete previous cache
# -t will allow accessing docker image using terminal
docker build --no-cache -t bp .
# step 04: Run the docker image
# allow terminal using -ti
# allow port using -p
docker run -it -p 8888:8888 bp
# if this gives error, try another port
docker run -it -p 8889:8889 bp
# If you have already running this container
# docker ps
# docker stop container_id_obtained_from_docker_ps
# this will open python, do not close it.
# open the new tab on the terminal to go inside the container.
# step 05: Go to docker terminal and run commands there
docker ps
docker exec -ti CONTAINER_ID_from_docker_ps /bin/sh
ls
pwd
# run python script
cd /home/bp/docs/scripts;clear;ls
which python
python --version
python example_json.py
# run jupyter notebook
cd /home/bp/docs/notebooks
# open the second link in the browser, where we can run the notebook.
# hit ctrl c to exit the python
# close the terminal tab where docker image was running.
# step 06: stop and remove the container
# first create new tab on the terminal
# hit ctrl d to close the running python
docker ps
docker stop CONTAINER_ID_from_docker_ps
docker rmi -f bp
Now close the docker app on your machine.This is my personal library intended to be used by only me. It's not a public library.