Recommendation engine for DevHub articles
Pre-work: save a TSV file downloaded from GA query explorer demo/tools https://ga-dev-tools.appspot.com/query-explorer/ set the following variables: metrics - ga:uniquePageviews dimensions - ga:pagePath and whatever the dimension for your custom user ID is Copy env.tmpl to .env and put in your values
parseTsvToCsv.py - step 1, cleans/prepares data. input:tsv file, output: csv file with the same name writeDataToMongo.py - step 2, takes cleaned up csv file and imports it into the pageviews collection. This collection has a record for each user with an array of pages they visited. New data is merged with existing data and de-duped associate.py - step 3, does the associative work - users that went to X page also went to Y page. Stores data in the associations collection. Calculated new each time, drops old data. getAssociations.py - Using information from the associations collection, make an array for each URL that has associations
compareAssociations.py - Compare the associations collection with the related collection, display differences. Data is not stored anywhere or changed; this allows us to easily see the added value of this process.