A quick test of a Viz visualisation library. It's purpose was to establish if displaying document data in graph structures can be used in front end. To run, open index.html in browser.
Experiments on data extraction from PDFs containing tables of numerical data. Using a number of different techniques to clean up the data.
analyseTable.py- extract data from a specific area of a specific pagefindTables.py- find all tables on a specific pageextractTableData.py- find tables on all pages and build a list of all of the metrics uses (net revenue, costs, headcont, etc.)
You can run the scripts without any arguments.
Check the diff between two PowerPoint presentations. Returns a list of diffs by page.
An neural nets experiment in Bitcoin price prediction. To build a model, create Book1.csv with prices in column A. This will create a model trainingData. Once model exists, the script will use lastDay.csv to get data for classification. From that, the precition is made.