dlumian / programminglanguagetrends

Exploratory analysis of Kaggle dataset on programming languages.

ProgrammingLanguageTrends

Exploratory analysis using Kaggle dataset on programming languages.

Primary forecasting model is Prophet.

Folder Structure

• data
  • archive.zip (zipped data which should be extracted and renamed to data.csv in the data folder)
  • data.csv
    • Week: Object type with date data in month/day/year format
    • Python/Java/C++: Cols with search count of that language for each week
• environment
  • environment.yml: list of all required packages
• imgs
  • Location for images generated to be stored
  • img paths are used in this readme, so moving or renaming them may break links
• notebooks
  • InitialEDA.ipynb: data investigation and summary statistics
  • ProphetModels.ipynb: primary analysis

Data Summary

Data summary generated by Skimpy package.

Forecast

For Python, Java, and C++ training data of counts of internet searches were used to predict future trends with Prophet. Counts were collected weekly from 2019-04-21 to 2024-04-21. Forecasts are projected weekly for the following year (52 predictions).

For each graph, a vertical dashed line indicates the end of existing data and the transition to predictions.

NOTE: The bottom bar is a slider which can be used in interactive plots to zoom in on a subsection of the data.

Python

Java

C++