Tuesday, April 3, 2018

Project Overview

Upon my graduation this May, I will be working at Progressive Insurance as a data analyst. From what I know, I will be working with insurance claims data in some capacity.

Throughout the next three posts, I will be building a Python program that does the following. This small program will simulate a large scale program that could be used by a major insurance company to be run daily and send out emails to claimants with the information for the claim they filed that day.

  1. The program will read a csv file ClaimsToday with claims information. Every row will hold information for a different claim. In reality, for a big insurance company, there would be thousands of rows, but my file will contain a small sample of claims. The various columns will hold the different pieces of information for each claim, including claim number, claimant first name, claimant last name, address, phone number, email, description, damage, location of accident, date of accident, time of accident, police involvement etc.
  2. Then the program will check the data for completeness and validity. Once the data is read into the program, we want to make sure that it is clean. Certain required fields must be filled, such as claim number, first and last name, and email. Other fields will have to be checked for formatting, for instance, if phone number is not a number or if claim number is not in the right format (we'll set the standard for a claim number to be six digits for the purpose of this exercise), the program should produce an error message to report the deficiencies in the data. Note that large corporations have entire departments dedicated to data cleansing, so we will be doing a very small example of this sort of work here.
  3. Finally, if the data is clean, the program will format a message with the information for each claim, which will be automatically sent to the email associated with the claim. I will be sending all emails to my personal email account for the purposes of testing the program.
My next blog post will cover the first step of this process, reading the data, so stay tuned!

Fun fact, I was Flo from Progressive this past Halloween! Until next time!





No comments:

Post a Comment