Daily Run Sheet is not a daily manual work anymore
Automated daily run sheet was a problem that I solved for my company when I was a part of it.
Note: I might not have a working example to show this, given it was my work done in my previous organisation
The entire algorithm basically consisted of 3 major steps:
- Address structuring
- Identification of addresses
The addresss structuring is a huge problem given addresses of India which are highly unorganised. What adds to a Data Scientist’s pain is the address field being a free text for customers, you can see all kinds of radomness there! Hence, cleaning of the address data is utterly important and no NLP libraries built for english languages can be of major help here, since addresses have their own stop words. Training of this address data is another important step to be able to build a model with accuracy.
Once the data is trained well, choosing the right algorithm that can help identify each address is another major challenge. Once that’s achieved with enough confidence, all you need to do is use a clustering method to cluster your addresses.
I can talk in detail about each of the leg of the problem and we solved it during the session.
Imagine if the process of creating a daily run sheet every morning for field executives has been automated to all levels. A solution that can automatically:
- find out which sub-locality each package belongs to
- how many FEs are available for a particular DC on a day
- allot packages to these FEs
- decide the order of delivery of each package in a route
Basic data science knowledge
A computer science engineer turned decision scientist turned data scientist. I have an experience of ~4 years.
Having worked closely with the board of directors of 3 startups in India & Indonesia, I am known for my business understanding, problem solving approach and obviously driving data science problems to the final execution.
a yoga lover, a poetess, painter, avid trekker & wanderer who is best at talking to people and learning about them