Text Analyzer Pro
Private, analyst-friendly text clustering for spreadsheet data.
- Passing tests
- 187
- Clustering models
- 4
- Privacy model
- Local
Data & BI AnalystKolkata, India
Aneek
Hait
I do the analysis, and build the tool when the tool doesn’t exist.
Officially: Data & BI Analyst at Accenture in Kolkata. In practice it’s about half analysis and half building the thing that makes the analysis possible — waiting two sprints for a script is how a question goes stale.
I let a model write the first draft of a lot of that code. It doesn’t get to merge itself.
Now
Building reporting systems at Accenture, shipping small tools that remove my own bottlenecks, and looking for work where an analyst is allowed to build.
01 — Work
Things I built because they bugged me.
Started as things I needed, couldn’t find, and decided to write instead. Every figure comes from the repo it links to.
TicketAudit
Offline desktop tool for auditing ITSM ticket exports — handles 100k–300k row files that choke Excel.
- Passing tests
- ~640
- Rows in 8.4 s
- 300k
- Languages detected
- 33
TicketLens
Privacy-first desktop workbench for IT ticket intelligence — local embeddings, semantic clustering, and LLM-powered insights without data leaving your machine.
- Passing tests
- 498
- Analysis modules
- 10
- Offline & private
- 100%
- Compute backends
- 4
ITSM Ticket Intelligence Command Center
Executive operations dashboard evaluating incident queue velocity, reassignment friction, CSAT benchmarks, and shift-left automation ROI.
- Incidents modeled
- 20,000
- Avg CSAT rating
- 4.21
- Reassignment hop levels
- 5+
- Shift-left deflection ROI
- $2.4M
02 — Analysis & studies
Pushing past the obvious answer.
Methodological studies and statistical evaluations — intervals, effect sizes, and evidence over gut feel.
IMF PortWatch — Global Maritime Disruption Intelligence
Live macro supply chain intelligence tracking daily vessel transit disruptions and rerouting economics across all 28 strategic global maritime chokepoints.
- Daily transits analyzed
- 78,372
- Strategic chokepoints
- 28
- Cape of Good Hope surge
- +98%
- Suez Canal volume drop
- -62%
Titanic Survival Analysis
A familiar dataset pushed beyond familiar conclusions.
- Passengers studied
- 1,309
- Survival odds ratio
- 11.3×
- Passing tests
- 49
03 — Side projects
Things I built outside the day job.
Independent experiments and useful tools, separate from client work.
Atomic Focus
Free AI Pomodoro timer that breaks overwhelming tasks into atomic steps and blends ambient sounds for deep work.
- Public access
- Free
- Ambient soundscapes
- 6+
- Session storage
- Local
04 — Experience
How I ended up here.
Started in delivery, moved toward the decision. The build instinct never went away.
08/2023 — Present
Software Engineer (Data Analyst)
Accenture / Kolkata
- Analyzed 100+ structured and unstructured datasets containing millions of data points.
- Built Tableau and Excel reporting that surfaces trends, outliers, and business priorities.
- Translated text analysis and descriptive findings for marketing, finance, product, and senior stakeholders.
07/2022 — 08/2023
Associate Software Engineer (BI Reporting)
Accenture / Kolkata
- Prepared and validated dashboard-ready datasets for recurring stakeholder reporting.
- Improved reporting quality through consistency checks, issue resolution, and cleaner preparation routines.
03/2022 — 05/2022
Software Developer Intern
Accenture
- Built software delivery foundations across Agile, testing, automation, Selenium, and Java.
05 — Toolkit
What’s on the desk.
Grouped by what I’m doing rather than by vendor, because the question picks the tool.
Analyse
- SQL
- Python
- pandas
- Statistical testing
- Confidence intervals
- Data QA
Build
- PySide6
- scikit-learn
- CLI + GUI tooling
- pytest
- Workflow automation
Ship & explain
- Power BI
- Tableau
- Excel
- Dashboard design
- Stakeholder reporting
How I work
- AI-assisted development
- Reproducible analysis
- Tests on anything that ships
- Cross-functional delivery
06 — Notes
Things I got tired of explaining twice.
Short pieces on the statistics and habits behind the analysis.
What an analyst learns by shipping a desktop app
Building Text Analyzer Pro changed how I scope analysis requests more than it changed how I write Python.
The most-analysed dataset is still misread
Everyone knows the Titanic headline. Far fewer have looked at what happens when you stratify class and sex jointly instead of marginally.
07 — Contact
Bring me
the messy one.
Fuzzy question, dubious data, no obvious owner. I’m interested in analysis, BI, and reporting work where an analyst is allowed to build the tooling as well as read the output.