Skip to content
AS / SIGNALS

Arnav Sahai · Data · Risk · Cybersecurity

I turn data, cyber risk, and policy evidence into decisions a skeptic can inspect.

Columbia SIPA MPA (Data Science for Policy · Cybersecurity). Former Accenture. Open to data, risk, and cybersecurity analyst roles.

Signal graphillustrative
Connected analyst signals: data, risk, and cybersecurity feeding a decision nodeDataRiskCyberPolicyDecide

Hover or focus a node

About

Who I am, and how I work with data and risk

I am a Columbia SIPA MPA graduate specializing in Data Science for Policy and Cybersecurity. My work sits at the intersection of analytical rigor, technical systems, and public-interest decision-making.

Professional headshot of Arnav Sahai

New York, NY

Education

  • Sep 2024 - May 2026

    Master of Public Administration

    Columbia University — School of International and Public Affairs

    Data Science for Policy; Finance & Economic Policy; Cybersecurity coursework. Degree conferred May 20, 2026.

  • Aug 2015 - May 2019

    Bachelor of Technology, Computer Science (Engineering)

    Guru Gobind Singh Indraprastha University

    Computer science engineering at Maharaja Surajmal Institute of Technology, affiliated to Guru Gobind Singh Indraprastha University.

Experience

Consulting, analytics, and delivery roles. Impact-oriented bullets from verified resume content, without invented metrics.

  1. Community Planning Fellow

    Sep 2025 - Apr 2026

    Fund for the City of New York — Manhattan Community Board 9 / New York, NY

    • Completed a Community Planning Fellowship placed with Manhattan Community Board 9, conducting applied research on housing affordability and equity in West Harlem.
    • Used ACS data, GIS, econometrics, impact measurement, and a service saturation analysis to show how affordability gaps and supportive-housing placement vary across the district.
    • Produced board-facing deliverables, including a final presentation and resource guide, and translated the analysis into a decision framework for CB9 leadership.
    Policy analysisGISEconometricsImpact measurementStakeholder research

    Related work: Manhattan CB9 Housing & Supportive Resources Analysis

  2. Research Consultant

    Sep 2025 - Dec 2025

    Sinclair Capital / New York, NY

    • Served as a client-facing consultant, developing an advanced statistical and financial model to investigate a dynamic business question and inform decision-making.
    • Structured an ambiguous problem into an executable analysis and communicated recommendations to a non-technical audience through written and oral presentation.
    • Produced well-documented model assumptions, inputs, and outputs to support reproducibility and stakeholder review in a fast-turnaround setting.
    Statistical modelingFinancial analysisClient consultingExecutive communication
  3. Development Impact Measurement Intern

    Oct 2025 - Dec 2025

    International Finance Corporation (World Bank Group) / New York, NY

    • Applied statistical and quantitative modeling and data mining across 15+ projects, designing custom analyses and benchmarking outcomes against sector metrics to support development and financial performance interpretation.
    • Owned end-to-end data management: cleaned and validated complex datasets, built relational databases, and maintained methods/assumptions and metadata documentation for reuse and auditability.
    • Implemented data quality assurance checks and repeatable SQL/Python/R workflows to improve integrity of inputs feeding analytical models and reporting for senior stakeholders.
    PythonRSQLData qualityImpact measurementRelational databases
  4. Associate Board Member

    Jul 2025 - May 2026

    BetaNYC / New York, NY

    • Guided and mentored participants in civic-tech engagement workshops.
    • Coordinated cross-sector initiatives across residents, agencies, and nonprofits.
    Civic technologyMentoringCross-sector coordination
  5. IT Consultant

    Oct 2024 - May 2026

    Columbia University / New York, NY

    • Supported 500+ students and faculty, resolving access, account, and endpoint issues under deadline with accurate documentation.
    IT supportDocumentationEndpoint troubleshooting
  6. AI Policy Research Intern

    Jan 2024 - Jun 2024

    Global Governance Initiative / New Delhi, India

    • Researched technology, AI governance, and industry trends and translated complex analysis into clear, decision-oriented recommendations for stakeholders.
    AI policyTechnology researchStakeholder communication
  7. Application Development Associate (Pharmaceutical client — Bristol Myers Squibb)

    Oct 2019 - May 2021

    Accenture / Bengaluru, India

    • Supported production data systems for a global pharmaceutical company, building data-integration processes (Java, Spring MVC) and monitoring data across 1,500 interfaces with Splunk to detect anomalies and quality issues.
    • Performed production support and defect triage, contributing to stable data flows and timely resolution of issues impacting downstream consumers.
    • Coordinated with cross-functional and global teams in an Agile environment, translating business requirements into technical updates for non-technical partners.
    JavaSpring MVCSplunkData integrationProduction supportAgile
  8. Data Analyst Intern

    Mar 2019 - May 2019

    The Smart Cube / Noida, India

    • Conducted procurement and supply-market research, benchmarking suppliers and analyzing cost trends and drivers to inform client decisions.
    • Designed custom analyses and Tableau/Power BI reporting and improved data accuracy by 70% through redesigned ETL and QA workflows.
    TableauPower BIETLQAMarket research

Letter of recommendation

  • Recommendation / March 12, 2026

    Letter of Recommendation — Fund for the City of New York

    Recommendation from Dr. Aldrin Rafael Bonilla, Executive Vice President of the Fund for the City of New York, covering the Community Planning Fellowship with Manhattan Community Board 9 (September 2025 to April 2026).

    Role: Community Planning Fellow

Projects

Selected projects

Cleaning, modeling, event studies, and papers with written findings.

8 items

Project

Data Breach Disclosures & Equity Market Reactions

Role
Analyst / researcher (reproducible R project)
What I did
Event-study and difference-in-differences analysis of how U.S. equity markets react around corporate data-breach disclosures.
Findings
  • Documented run (2026-04-20): mean CAR(0,+10) ? ?0.12% (n = 224 non-confounded); mean AR sharply negative at disclosure day (t = 0).
  • CS/DID overall dynamic ATT ? ?0.22 (95% CI [?0.42, ?0.02]); pre-trend p ? 0.91.
  • R
  • tidyquant
  • Event study methods
  • Callaway- Sant- Anna / did

Project

Elon Musk Tweets & Tesla Stock — Predictive Modeling

Role
Analyst
What I did
Sentiment, lag, ARIMA/ARIMAX, and causal-impact analysis of Elon Musk tweets (2010–2025) against Tesla prices, finding a statistically detectable but weak relationship (r ≈ 0.16).
Findings
  • Daily tweet sentiment and Tesla adjusted price were weakly positively correlated (r ≈ 0.156 contemporaneous; 0.156 at 1-day lag; 0.155 at 2-day lag).
  • The relationship is statistically detectable but limited; market conditions and fundamentals likely dominate price moves.
  • R
  • tidytext (Bing lexicon)
  • quantmod
  • ARIMA / ARIMAX

Project

Blink to Text

Role
Co-author (B.Tech minor project team, Maharaja Surajmal Institute of Technology)
What I did
Real-time blink-detection application that converts eye blinks into text so people with Locked-in Syndrome can communicate independently on a low-end computer and webcam.
Findings
  • Selected HOG + SVM over Viola-Jones because it is more robust to non-frontal pose and illumination for users who cannot stay facing a camera.
  • Designed the system so it can be started, paused, and operated entirely with blinks, reducing dependence on caretakers for communication.
  • OpenCV
  • dlib
  • HOG + SVM
  • Haar Cascade

Project

Russian Troll Twitter Activity Classifier

Role
Analyst (Columbia SIPA — Text as Data)
What I did
Text-as-data analysis of IRA troll tweets: dictionary methods, sentiment/emotion profiling, and a Naive Bayes classifier that outperformed a classification tree (71.9% vs. 50% accuracy).
Findings
  • RightTroll accounts used election keywords more, with a spike in late 2016 around the U.S. presidential election.
  • LeftTroll tweets had higher average sentiment and more profanity; RightTroll tweets had slightly negative average sentiment.
  • R
  • quanteda
  • sentimentr
  • caret

Project

Brooklyn Fare-Evasion Arrests — Microdata Cleaning & Description

Role
Analyst (Columbia SIPA DSPC 7514)
What I did
Harmonized 2016 Brooklyn Defender Services and Legal Aid Society fare-evasion arrest files, recoded race/ethnicity, and mapped enforcement concentration across subway stations.
Findings
  • In the BDS file, Non-Hispanic Black clients were the largest recoded group (1,359), followed by Hispanic (493) and Non-Hispanic White (165).
  • The sample represents BDS/LAS clients, not all Brooklyn fare-evasion arrests; dismissal is observed only in LAS data.
  1. Raw files
    42
  2. Join keys
    55
  3. Race recode
    71
  4. Missingness rules
    78
  5. Documented sample
    86

Illustrative confidence index

  • R
  • tidyverse
  • fastDummies
  • Administrative microdata cleaning

Project

ASAN MACRO — AI-Powered Trade Intelligence Agent

Role
Sole builder (data layer, agent, evaluation, and reporting)
What I did
Tool-augmented LLM agent that converts structured trade data into grounded, decision-ready economic sentiment reports.
Findings
  • Built an end-to-end agentic product for BRICS and sector-focused trade analysis with documented evaluation scaffolding.
  • Public demo deployed at asan-macro.vercel.app for trade sentiment briefs grounded in UN Comtrade, World Bank, and IMF series.
  • Python
  • SQLite
  • LLM tool-calling / RAG
  • UN Comtrade ingest scripts

Project

Caste and Household Expenditure Gaps in Uttar Pradesh

Role
Analyst (Columbia SIPA U7514)
What I did
ACCESS 2018 analysis of 3,002 Uttar Pradesh respondents showing a ₹1,385 mean monthly expenditure gap between General and Reserved castes that narrows but persists among landowners.
Findings
  • General-caste mean monthly expenditure ₹7,861 vs. ₹6,475 for Reserved castes (gap ₹1,385).
  • General-caste respondents are 24% of the UP sample but 33% of own-land agricultural households.
  • R
  • dplyr
  • DescTools
  • Household survey analysis

Project

Egg Price Fluctuations & Bird Flu — Forecasting

Role
Analyst
What I did
Regression and time-series analysis linking H5N1 outbreak timing to Grade A egg price surges and related equity responses.
Findings
  • Identified a 2- 3 month lag between H5N1 outbreaks and Grade A egg price surges.
  • Documented equity responses in the analysis for Cal-Maine (+18%) and Vital Farms (+12%).
  • R
  • Regression
  • Time-series analysis

Project Demo

Live demos you can open and walk through

Published project sites. Demos stay unloaded until you choose to run them.

4 live demos

Manhattan CB9 Housing & Supportive Resources Analysis

Data-driven housing affordability and Fair Share analysis for Manhattan Community Board 9, including a revised decision framework for project evaluation.

arnav504.github.io/CB9_housing/Open

The live demo is a third-party site. It is not loaded until you ask, so this page stays fast and private by default.

Open in new tab

Cyber Narrative Radar

Full-stack early-warning system that turns public cyber discourse into explainable alerts, organization drilldowns, and narrative clusters for analyst workflows.

cyber-narrative-radar-web.onrender.com/Open

The live demo is a third-party site. It is not loaded until you ask, so this page stays fast and private by default.

Open in new tab

ASAN MACRO — AI-Powered Trade Intelligence Agent

Tool-augmented LLM agent that converts structured trade data into grounded, decision-ready economic sentiment reports.

asan-macro.vercel.appOpen

The live demo is a third-party site. It is not loaded until you ask, so this page stays fast and private by default.

Open in new tab

Escalation Risk Lab

Interactive dashboard comparing AI-assisted vs. human-only crisis escalation risk across synthetic cyber, geopolitical, and energy incidents.

crisis-analyzer-ai.lovable.appOpen

The live demo is a third-party site. It is not loaded until you ask, so this page stays fast and private by default.

Open in new tab

Reports

Reports, memos, and research papers

Quantitative papers and analysis PDFs: cleaning, modeling, and written findings.

7 items

  • Research PaperPDF

    The Impact of Elon Musk's Tweets on Tesla's Stock Price

    Sentiment, lag, ARIMA/ARIMAX, and causal-impact study of Musk tweets (2010–2025) and Tesla prices, finding a weak positive correlation (r ≈ 0.16).

    Role: Author

  • Research PaperPDFMarch 2025

    Russian IRA Troll Tweets — Dictionary Methods, Sentiment, and Classification

    Text-as-data analysis of LeftTroll vs. RightTroll IRA accounts, including election-keyword dictionaries, sentiment/emotion profiling, and a Naive Bayes classifier (71.9% accuracy).

    Role: Author

  • ReportPDFFebruary 2026

    Brooklyn Subway Fare-Evasion Arrests — Microdata Cleaning

    Harmonized 2016 BDS and LAS public-defender arrest files, recoded race/ethnicity, and described station-level enforcement concentration.

    Role: Author

  • ReportPDFJanuary 2026

    Caste and Household Expenditure Gaps in Uttar Pradesh (ACCESS 2018)

    Survey analysis of 3,002 Uttar Pradesh respondents: General-caste mean monthly expenditure exceeds Reserved-caste means by ₹1,385, with a smaller gap remaining among landowners.

    Role: Author

  • Research PaperPDFFall 2025

    Quantifying the Impact of Extreme Weather Insurance Costs: The LA Wildfires Case Study

    Columbia SIPA practicum co-authored study treating the 2025 Los Angeles wildfires as a natural experiment for insurance-sector transmission of climate-driven systemic costs.

    Role: Co-author (practicum team)

  • Research PaperPDF2018

    Blink to Text: A Real-Time Blink Detection and Text Displaying Application

    Team paper on a webcam-based application that detects blinks in real time and converts them into text so people with Locked-in Syndrome can communicate independently.

    Role: Co-author

  • ReportPDFDecember 2018

    Blink to Text — B.Tech Minor Project Report

    December 2018 minor project report submitted to Maharaja Surajmal Institute of Technology (GGSIPU) for the Blink to Text computer-vision application.

    Role: Co-author

Certifications

All certificates, ready to view on this page

Certificates from the same organization are grouped together. Open a document to view it here.

Professional certifications

Google / Coursera1 certificate

  • Feb 2026PDF
    Google Cybersecurity Professional Certificate

    Nine-course professional certificate. Individual course certificates are listed below.

    • Foundations of Cybersecurity

    • Play It Safe: Manage Security Risks

    • Connect and Protect: Networks and Network Security

    • Tools of the Trade: Linux and SQL

    • Assets, Threats, and Vulnerabilities

    • Sound the Alarm: Detection and Response

    • Automate Cybersecurity Tasks with Python

    • Accelerate Your Job Search with AI

Anthropic5 certificates

  • May 2026PDF
    Claude 101
  • May 2026PDF
    Claude Code 101
  • May 2026PDF
    AI Fluency: AI Capabilities & Limitations
  • Mar 2026PDF
    AI Fluency: Framework & Foundations
  • Mar 2026PDF
    AI Fluency for Students

Columbia University2 certificates

  • Jun 2026PDF
    Prompt Engineering & Programming with OpenAI
  • Jun 2026PDF
    Building Customized LLMs with OpenAI

Bloomberg2 certificates

  • May 2026PDF
    Bloomberg Finance Fundamentals
  • PDF
    Bloomberg Spreadsheet Analysis (BQL)

DataCamp3 certificates

  • Mar 2025PDF
    Python Programming Track
  • Mar 2025PDF
    Data Manipulation with Python Track
  • Mar 2025PDF
    Importing & Cleaning Data with Python Track

IBM1 certificate

  • Mar 2020PDF
    IBM Data Science Foundations — Level 1

Simplilearn1 certificate

  • Mar 2020PDF
    Statistics Essential for Data Science

McKinsey1 certificate

  • Jun 2026PDF
    McKinsey Forward

Leadership and community

Global Governance Initiative1 certificate

  • PDF
    YLP Project — Ethical AI in Autonomous Vehicles

    Certificate of participation for work on ethical AI decision-making in autonomous vehicles.

TED1 certificate

  • Mar 2018PDF
    TEDxDabri Certificate of Appreciation

Maharaja Surajmal Institute of Technology1 certificate

  • 2018PDF
    Chief Editor — VISION 2018

    Certificate of appreciation for completing work as Chief Editor of the college magazine.

Human Circle1 certificate

  • Oct 2019PDF
    Young India Challenge — Volunteer

    Certificate of excellence for volunteering at Young India Challenge: Sustainable Living, New Delhi.

Analyst workspace

From narrative to action.

A front-end walkthrough of how unstructured threat text becomes a brief. All records are fictional and labeled illustrative.

Illustrative

Raw narrative

Several municipal utilities in the Great Lakes region reported overlapping phishing lures this week impersonating a billing portal. A low-volume ransomware leak site named a water-sector contractor, without proof of access. Local journalists repeated the claim. No confirmed OT disruption. Timing overlaps a regional storm that is already stretching operator attention.

Notes

How I think on the page

Short writing on narrative intelligence, risk communication, and decision-ready analysis. Process and findings, not a tool list.

Contact me

Email, LinkedIn, and GitHub

Reach me directly by email, or open LinkedIn and GitHub.

arnav.sahai@columbia.edu

New York, NY