Machine learning for General Ledger provide a complete Data Science framework for Finance Data Analytics. Scope of this package includes Finance General Ledger, sub-ledgers, accounting entries analytics and complex Data Science operations using The Julia language.
This package/presentation is meant for ERP consultants, IT Developers, Finance, Supply chain, HR & CRM managers, executive leaders, or anyone curious to implement data science concepts in ERP apps for small, medium & large Org.
ERPs are backbone systems to any living organization, whether small, medium, or large. Almost every organization use ERP concepts, whether in a piece of paper, excel file, MS Access DB or large sophisticated systems like SAP, Oracle data warehouse & databases.
Organizations deal with many ERP modules like Buy to Pay, Order to Cash and GL to manage Finance, Supply chain operations daily.
ERP systems aren’t luxury apps anymore, Often Small businesses see GL (or perhaps entire ERP system), intended only for Big organizations. However, even Big Org operate in small units and merge later into a Corporate Ledger.
Having a system to read, write and understand your financials books brings extraordinary benefits to any size of business.
Recent popularity of software like Quick-books, NetSuite, Tally are proof that ERP are the way to manage business successfully. People use piece of papers, notebooks, excels, word documents in their daily life to manage expenses, record Actuals, plan future Budgets and even forecast future savings to some extent.
Medium, Big Org just happen to use sophisticated ERP systems like SAP, Oracle, PeopleSoft, Coupa etc. but conceptually, they don’t do anything different other than following simple ERP concepts daily to manage critical business operations, record Actual and forecast earnings.
Problem Statement: IT industry & Giant Market leaders created extraordinary ERP systems to record transactions. Even a small merchant, now uses his mobile to take picture of invoice and pays vendor online. Medium organizations can auto-correct invoices and process vouchers, and Big Org’s ERPs can count live inventory across regions. These systems provide a unified platform for all small, medium and large entities to record i.e. write their data in structured or unstructured forms into their secured on-premise or cloud locations.
Read, Write and Understand are 3 different aspects of ERP data, while these ERP systems achieved perfection in “write”, there is a lot more to do in “read and understand data”.
This Julia Language GeneralLedger.jl is built to serve this purpose only.
This package will provide a complete Analytic Software environment, which can be deployed as a bolt-on or independent app for all data extract, load, transformation, ad-hoc reporting & analytics, visualizations and tooling to support Data Science using AI, ML predictive Analytics.
Why Julia Language:- because it take a village to create Analytics platform, Julia is not solving 2 language problem here in ERP space, ERP systems already have 4, 5 or 6 language problems.
Transactions systems are not primarily used for Reporting/analytics purpose, imagine using Walmart pay register to run data analytics.
Organizations spend millions of $$s to move their data to appropriate data platform, they use ELT, ETL, create Data warehouses, Data lakes to keep this data in sync.
ERP systems run complex data operations to clean this data set before it’s made available to business users, supply chain managers and executive leaders in form of analytics tools like Cognos Data marts, Tableau, Power BI, Oracle OACs, Kibana or Google Analytics like business intelligence tools.
While these tools are great for operational reporting, they lack analytics.
I am not talking about simple data operations like removing duplicates, drill down, drill through reporting or creating bar- graphs, drawing moving average lines on simple graphs.
By Data Science, I mean, plotting entire company General Ledger data on a single graph. I am talking about using Network science/ graph theory to manage supply chain, using Machine Learning prediction for Real time inventory use charts.
Julia solves this problem, it’s one Data Science language, which can be used for data loading, cleansing, transformations, visualizations and then running advance AI, ML or Deep learning computational operations on huge data sets.
Real time, real life statistics:- “Walk Like Python; Run Like C” is enough to pick this language for all my daily work. ERP leaders have a business to run, and a community to serve.
“Julia is high performance”, made for Asynchronous, parallel, distributed GPU computing,
Julia language provides fast computations for large data sets and great assets for Statistical programming, makes Julia Language is THE language for ERP Analytics.
In this presentation, I will share my Julia notebooks for General Ledger. I will share, how Julia provide data structure for LEDGER, FINANCE using sophisticated TYPE system. How data can be quickly transformed to Fact/Dimension star schema for self-service and real-time reporting. and I will share some of my Julia notebooks I used to perform Time Series, Deep Learning and Graph analysis on Finance data.