AI data analysis & comparison of data -- 2

Description:

We are seeking a skilled developer to create an innovative tool that automates the analysis and comparison of insurance settlement files (typically in PDF form created by Xactimate- essentially pdfs of data which mimics spreadsheets; examples attached). The primary objective is to develop a system or ability that can accurately identify discrepancies in settlement amounts allocated room by room, as outlined in insurance settlement PDFs, against our own estimates. Typically an insurance company will provide an Xactimate PDF file to us that contains allotted monies for an insurance loss to estimate the cost of the loss. Our role is to produce our own version of that same file, based on our calculations and observations/costs which are typically way more than what the insurance company allows initially. The data tool therefore would analyze these files (broken down by location inside the property), and view, analyze, compare the allotted amounts and specific items allowed by the insurance company and the same location compare and contrast our opinion of what this should look like (usually way more money that the insurance company allows), and then point out in a table the deficiencies of the insurance company's version to ours, and point out the financial impact of the omissions on the part of the insurance company and/or calculations made by their file relative to ours. By way of example:

our file says 100 sq feet at $10 per square foot, their file say 80 sq ft at 8 dollars per foot. the table should then identify the room, identify that the insurance company allotted 80 sq ft and we calculated 100 sq ft, then our price of 10 and their price of 8 and the total economic difference of our estimate $100 dollars their estimate $64 dollars. etc..

then after the detailed table/analysis write a brief cover letter outlining the differences and speaking to them at a summary high level and highlight them.

Core Requirements:

PDF Data Extraction: Ability to extract and interpret data from various PDF formats, including scanned documents and text-based PDFs.

Data Standardization: Develop a methodology to standardize room names and item categories across different documents for accurate comparison.

Data Analysis: Implement a comparison mechanism to identify missing items, calculate differences in cost estimates, and highlight discrepancies between the insurance company's files and our estimates.

Result Presentation: Design a user-friendly interface or report format that clearly presents the findings, highlighting areas where the insurance settlements are deficient.

Preferred Qualifications:

Proficiency in programming languages suitable for data manipulation and analysis (e.g., Python, R).

Experience with OCR technologies and natural language processing.

Familiarity with machine learning platforms and data analysis libraries.

Ability to implement a flexible and scalable solution that can handle a large volume of documents.

Expected End Result:

The final product should be a tool or platform that provides a detailed comparison report, indicating discrepancies in the insurance settlement amounts versus our estimates. The report should:

Clearly identify and itemize each discrepancy in settlement amounts on a room-by-room basis.

Provide a summary of omitted items and under-valued estimates.

Offer a concise overview of potential areas for re-evaluation and additional claims.

Be exportable or easily shareable in formats such as CSV, PDF, or through a web interface.

Proposal Instructions:

Please provide a brief overview of your approach to this project, including:

Technologies and platforms you plan to use.

Estimated timeline and budget.

Any similar projects you have completed.

We are open to suggestions regarding the technology stack and approach, as long as it is efficient, cost-effective, and delivers accurate results. We value creativity and efficiency and are looking for a developer who can provide a comprehensive solution to this challenge.

An illustration and description of some challenges:

typically each file is broken down by room name and various costs/items associated with the estimated cost for repair in those rooms (see attached file example). One issue is that our file we create may use a different naming convention that the one supplied by the insurance carrier. So we first want to map and associate the rooms to each other in both files and make sure the AI is comparing the line items in the associated room with the correct associated room in both files. So the AI should ask if den in one file = study in another (as an example). Then detail the differences and or point out the total omission of a room (if that has happened) in the insurance companies file.

here is an excerpt and example of the data in the files.

Essentially the AI would have to match :

CAT #/item QTY ACT Description File "A" File "B" Difference Notation

34 walls 16 sq ft Two coats plaster 1/2" gypsum $237.02 $432.02 +$194.00 quantity the same $ less

Python Machine Learning (ML) OCR Ngôn ngữ lập trình R

ID dự án: #37909063

Về dự án