The application runs on localhost:5000 by executing python -m flask run while in the source folder.
The commit relating to first access of the github API: efc861c76652afacc6b0802f1f129cc7f372268a.
The basic idea is to demonstate the correlation between companies and the languages used using information gained from the top repos and top users on the site.
Backend - Flask and Python. Frontend - HTML, Javascript, CSS, and Jinja2 (a templating language).
Step 1: Through the Github API I get 300 of the top rated Github Repositories by number of stars and 300 of the top users by number of followers.
Step 2 - Languages: Dealing with the Repo: I create a json file that has a key being the language and the value being the number of repos that primarily use that language.
Dealing with the User: This was slightly more difficult to achieve. Again through the Github API I receive a slice of 10 of their repos and get the most used language within those 10. I then again create a json file containing key of language and value being number of users that most use that language.
Step 3 - Companies: Using a process similar to the one above, I get the the company affiliated with a repo/user and then the number of other repos/users that associate with that company creating two seperate key value json files. One for user and one for repo.
Step 4 - Combining: Once I had the four seperate json files I then combined them into two seperate json files one that had the companies combined and then one that had the languages combined. This is so as I am able to provide breakdown further through the use of pie charts.
Step 5 - Correlation: Using the top repos and top companies I create a json file that contains the company the user/repo is associated with as the key and then the value being the number of times a language is associated with a repo/is the top language for a user. eg. "freeCodeCamp": {"JavaScript": 1}, "facebook": {"JavaScript": 6, "null": 1, "Python": 1}
Step 1: Read in the corrCompanyLanguages json file.
Step 2: I get the top X number of languages by mention in the file and the top X number of companies by language use variety. The default being X = 10.
Step 3:
Using this information I am able to create the flow matrix showing the flow between company and language.
Example of flow matrix with 3 values for companies (A, B, C) and 3 values (X, Y, Z) for languages.
A | B | C | X | Y | Z
A 0 | 0 | 0 | 3 | 5 | 6
B 0 | 0 | 0 | 1 | 0 | 7
C 0 | 0 | 0 | 2 | 8 | 1
X 3 | 1 | 2 | 0 | 0 | 0
Y 5 | 0 | 6 | 0 | 0 | 0
Z 6 | 7 | 1 | 0 | 0 | 0
Flow being -
A -> X = 3
A -> Y = 5
A -> Z = 6
B -> X = 1
B -> Y = 0
B -> Z = 7
C -> X = 2
C -> Y = 6
C -> Z = 1
Step 4: I set up the labels associated with the flow matrix and randomly generate a colour to go with the label.
Both arrays are sent to the front end to be dealt with and displayed a combination of Javascript, CSS and HTML.
Default display of correlation between 10 companies and 10 languages.
Dynamic display of any number of correlations between the numbers of 1 - 30.
Further breakdown provided by piecharts containing information gained through the creation of the chord chart.
