AI Keyword CSV Sorter
Budget: $250 – $750 USD
I have a collection of CSV files filled with text fields, and I need an AI-driven script that can sift through every record, detect the presence of predefined keywords, and then sort, tag, or export the matching rows for easy review.
Scope
• Input: CSV file containing raw text in one or more columns.
• Processing: the solution should read each file, compare the text against a customizable keyword list, and flag or separate the rows that meet the criteria.
• Output: a new CSV (or set of CSVs) that clearly identifies matched entries—ideally with a simple “Match” column or a separate file for positive hits.
Technical preferences
Python is my default choice thanks to its mature libraries (pandas, scikit-learn, spaCy), but I’m open to alternative stacks if you can justify clear advantages. A lightweight ML approach or traditional rule-based matching is fine—accuracy and speed matter more than model complexity.
Deliverables
1. Fully commented source code or notebook.
2. README explaining setup, configuration of the keyword list, and execution steps.
3. Sample run on test data proving correct detection.
Acceptance
I’ll consider the job done once the script processes a provided sample CSV, outputs the expected matches, and I can change the keyword list without touching the code.
If you have ideas on expanding later to sentiment or phrase-level matching, let me know, but for now it’s all about keyword accuracy.
Scope
• Input: CSV file containing raw text in one or more columns.
• Processing: the solution should read each file, compare the text against a customizable keyword list, and flag or separate the rows that meet the criteria.
• Output: a new CSV (or set of CSVs) that clearly identifies matched entries—ideally with a simple “Match” column or a separate file for positive hits.
Technical preferences
Python is my default choice thanks to its mature libraries (pandas, scikit-learn, spaCy), but I’m open to alternative stacks if you can justify clear advantages. A lightweight ML approach or traditional rule-based matching is fine—accuracy and speed matter more than model complexity.
Deliverables
1. Fully commented source code or notebook.
2. README explaining setup, configuration of the keyword list, and execution steps.
3. Sample run on test data proving correct detection.
Acceptance
I’ll consider the job done once the script processes a provided sample CSV, outputs the expected matches, and I can change the keyword list without touching the code.
If you have ideas on expanding later to sentiment or phrase-level matching, let me know, but for now it’s all about keyword accuracy.
Related categories:
PHP
Python
Data Processing
Web Scraping
Software Architecture
Machine Learning (ML)
Data Extraction
Data Analysis