eCommerce AI

about e-commerce.ai

eCommerce.ai is a complete end-to-end Data Science, Graph, AI/ML Analytics technology framework to support data gathering, mining, wrangling, analysis including visualization & performing AI/ML Analytics on data, with intent to understand, support and predict eCommerce operations to support daily business operations for small, medium and very large organizations.

eCommerce.ai is an official project submitted as Galp's Hackathon Retail 4.0 Hackathon project work.

Main source code/example notebooks are executed & included in documentation published under GitHub gh-pages branch. 

Best way to understand content on this project is to go through gh-pages branch

documentation source code
complete source code

Platform
Frontend: Julia 1.7.1
Backend: Oracle OCI Cloud, Oracle ADW (Autonomous data warehouse) | TigerGraph/Oracle Graph DB
Rest API: Julia, TGCloud RESTAPI
AI: Julia, FLUXml.ai, Oracle AutoML

Implementation approach

eCommerce.ai takes a methodological business workflow approach (follow data) to solve this challenge.

Step 1:

At first, a detail analysis (much of the work) is done to understand, define end-to-end source to pay, order to cash, procure to sell business operations.
You will see, tons of examples included in this project,
These examples resemble real life commercial good procurement to sales including payments, accruals, receiving and expenses etc..

Step 2:

Next, 3rd part IOT data like, local community events, holiday calendars, long weekends, weathers, climatic conditions, type of data is gathered.

Step 3:

Then all of this data is combined, cleaned and wrangled in a format which can be used in Analytics.

Step 4:

Then after, following Analytics is run and made available (in form of Jupyter | Pluto notebooks) for business operations, KPI Dashboards and Executive dashboards. These KPIs help business leadership take effective operational intelligence decisions.

Final deliverables

Ad-Hoc reports :      Simple data queries
Analytics:                Self service reporting, analytics & visualization
Advance Analytics:   would | could | should
Predictive Analytics: Train, test and predict KPIs
Real time Analytics:  Running analytics on real time data

Assets | Folder | File Structure

README.md:  start here

assets    notebooks:  These notebooks can run standalone | Docker container            
              Jupyter -> Standard notebooks                    
              Pluto -> reactive, auto run , real time data refresh    

sampleData: small sample datasets
                    please see, GitHub doesn't allow to upload big amounts of data.

sampleData jupyter notebook -> This Julia notebook can be used to to generate volume of commerce data.

docs/src :  complete source code with executed samples
docs/make.jl:   This file is used to generate HTML Documentation of code.
                 similar to Python Sphinx | readthedocs | Jupyter eBook

src:    actual Julia eCommerce package
          There isn't much here, because all source code is included & executed as in-line documentation. but this folder/file is required for project to compile.

GitHub gh-pages branch    This is main starting point of this project. start here

more info

Proposed Architecture

Business process workflow

ERD

Explainable AI