Open to full-time roles — Tucson, AZ

Divya Dhole Data Analyst

MS Data Science at University of Arizona. I turn real government data from BLS, EPA, Federal Reserve and the Census Bureau into analysis that answers questions people actually care about. I use real government APIs — BLS, CMS, EPA, FRED, FBI, FEMA — to build analysis that answers questions people actually care about.

13
Projects Built
5+
Live Dashboards
27
GitHub Repos
Live Census ACS / HUD / FRED 3x affordable 2022 peak: 6.1x Payment Burden 2023 34.7% vs 28% lender guideline 12 states in crisis tier Price-to-income ratio 2000–2023 · 3.9x → 6.1x peak · Sources: FRED + Census ACS + HUD
P16 · US Housing Affordability Crisis↗
Housing Affordability Analysis 2000–2023
Census ACS + HUD Fair Market Rents + FRED
Home prices rose 161% from 2000 to 2023 while incomes rose 85%. Price-to-income hit 6.1x in 2022. Monthly payment burden reached 34.7% of income in 2023 vs the 28% lender guideline. 12 states in crisis or severely unaffordable tier. Hawaii worst at 9.6x.
Census ACSHUDFREDSQLPython
GitHub Policy / Economics HI CA IA State Rankings 2022 Crisis >8x Severely unaffordable 6-8x Unaffordable 4-6x Moderate 3-4x Affordable <3x All 50 states ranked by price-to-income ratio · Iowa (3.0x) to Hawaii (9.6x)
P16 · State Rankings + SQL Analysis↗
50-State Affordability Rankings
Census ACS B19013_001E + HUD FMR + FRED MSPUS
Hawaii (9.6x) to Iowa (3.0x) — every state ranked by price-to-income ratio. 7 SQL queries with LAG() trends and RANK() window functions. 6-sheet Excel workbook. Why the Iowa exception exists and why Washington state has great sun but terrible solar economics.
LAG() / RANK()SQLitematplotlibExcel
Python✦
SQL✦
Power BI✦
BLS JOLTS API✦
Federal Reserve FRED✦
EPA AQS✦
Census Bureau✦
Machine Learning✦
Tableau✦
scikit-learn✦
AWS✦
GCP✦
Databricks✦
XGBoost✦
Python✦
SQL✦
Power BI✦
BLS JOLTS API✦
Federal Reserve FRED✦
EPA AQS✦
Census Bureau✦
Machine Learning✦
Tableau✦
scikit-learn✦
AWS✦
GCP✦
Databricks✦
XGBoost✦
Data Projects
Live Labor / BLS 3.0% PEAK 2019COVIDGreat Resignation2023
P05 · Labor Economics↗
The Great Resignation
BLS JOLTS API — Real Government Data
Quit rate peaked at 3.0% in Nov 2021, +19.7% above baseline. Workers behind inflation for 27 of 48 months. Welch t-test p<0.001, Cohen's d=3.82.
PythonSQLscipyBLS APISQLite
Live FRED / Economics 9.1% peak Inflation Wages Fed Rate
P07 · Economics↗
Are Workers Falling Behind?
Federal Reserve FRED + BLS CPI
CPI peaked at 9.1% in June 2022. Fed hiked from 0.05% to 5.33% in 18 months — fastest tightening cycle since Paul Volcker in the 1980s.
FRED APIBLSPythonSQL
Live EPA / Environment 2000 2019 2023 NAAQS PM2.5 down -42% since 2000 · NYC AQI 218 in 2023
P08 · Environment↗
Who Breathes Dirty Air?
US EPA Air Quality System (AQS)
PM2.5 down -42% since 2000. But 2023 Canadian wildfire smoke pushed NYC AQI to 218 — worst in modern US history.
EPA APIPythonSQLseaborn
Live Census / Retail 2019 2020 2021 2022 2023 $8.5T COVIDcrash
P06 · Retail / Economy↗
What America Buys
Census Bureau Monthly Retail Trade Survey
COVID crashed retail -23.2% in April 2020. Recovered in 4 months. E-commerce accelerated 5.5 years ahead of schedule. Total 2023: $8.5T.
Census APIPythonSQL
Healthcare / CMS 23.3% readmission Critical High Moderate Low AUC: 0.614 5-fold CV p < 0.001
P01 · Healthcare↗
Hospital Readmissions
CMS HRRP Benchmarks
23.3% readmission rate. Logistic regression CV AUC: 0.614. Uninsured patients +14% risk. 4-tier risk stratification model.
PythonMLSQLscikit-learn
Investigative / CMS Pharma Payments ($) Prescriptions +64.1% MORE if doctor was paid
P03 · Investigative↗
Follow the Money: Pharma
CMS Open Payments + Medicare Part D
Paid doctors prescribe +64.1% more. Spearman r=0.134, p<0.001. $10k+ recipients prescribe 2.7x more.
CMS DataPythonscipySQL
Healthcare / SQL Avg Door-to-Doc: 79.7 min · LWBS: 2.23%
P02 · Healthcare SQL↗
ER Wait Time Analysis
ACEP Benchmarks · 8,000 visits
Avg door-to-doc 79.7 min (target <60). LWBS rate 2.23%. Night shift adds +55% wait. 10 SQL window function queries.
SQLPythonExcelSQLite
Retail / EDA Monthly Revenue Seasonality Index 85 87 96 98 100 98 101 103 99 107 118 131 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec RFM Customer Segments Champion12% Loyal28% Potential35% At Risk18% Lost7% 50,000 transactions · 10 stores · $8.97M revenue
P04 · Retail EDA↗
Retail Sales Analysis
50,000 transactions · 10 stores · 3 years
Seasonality index, RFM customer segmentation using NTILE(), YoY growth with LAG(). $8.97M total revenue across 10 product categories.
PythonSQLExcelseaborn
Live DEA + CDC / Investigative Fentanyl dominates Rx Opioids Heroin Fentanyl (Wave 3) Cocaine/Meth 2015201720192021 — 109K deaths
P09 · Investigative↗
Follow the Pills: Opioid Crisis
DEA ARCOS + CDC WONDER — Real Government Data
109,680 overdose deaths in 2022 — 46% from fentanyl. West Virginia: 79.2 per 100,000. DEA ARCOS: 2.83 billion pills shipped at 2010 peak. Three waves mapped.
DEA ARCOSCDC WONDERPythonSQL
Live USDA / Equity Native 17.6% Black 10.8% Hispanic 9.4% Asian 6.2% White 5.2% Food Desert % MS NM LA AR WV AL OK TN KY
P10 · Food Equity↗
Where America Cannot Eat
USDA Food Access Research Atlas — Real Government Data
16.1M Americans live in food deserts. Black rate: 10.8% vs 5.2% White. Native Americans average 5.8 miles to nearest store. Mississippi worst at 11.8%.
USDA AtlasPythonSQLseaborn
Flagship Project — Healthcare Fraud Detection
Live App ML / Healthcare / Fraud Detection CRITICAL HIGH OIG Confirmation 98% 746 / 761 CRITICAL already confirmed fraud PR-AUC: 0.9996 Fraud score distribution · 50K providers · 761 CRITICAL · 98% OIG confirmed
P14-P18 · Flagship · 24+ commits↗
CMS Medicare Fraud Intelligence
CMS Part D 2022 + OIG Exclusions + XGBoost + Isolation Forest
98% of CRITICAL-tier flags confirmed on OIG exclusion list. XGBoost PR-AUC 0.9996 vs baseline 0.015. 43.5x fraud/normal flag separation. SHAP explainability, Streamlit provider lookup, state choropleth, 5-page dashboard. The model found them purely from billing pattern anomalies.
XGBoostIsolation ForestSHAPCMS Part DStreamlitSQL
Why This Gets You Hired
Healthcare fraud costs the US $60-90 billion per year. Every healthcare company, insurance company, and Medicaid contractor wants someone who can find it.

This model analyzed 50,000 real CMS providers and flagged 761 as CRITICAL. Of those, 746 were already on the federal exclusion list. The model found them without ever seeing the labels during training.

That sentence is a job offer.
View on GitHub Live App ↗
Live App Solar / Finance / Streamlit Payback 7.3yr 100kW · Arizona IRR 12.7% NPV $147K 30% ITC · MACRS Cumulative cash flow · Red = pre-payback · Green = positive
P14 · Solar ROI Calculator↗
Solar Project Financial Calculator
EIA Commercial Rates + NREL Solar Resource
Select a state → auto-fills real EIA utility rate + NREL capacity factor → instant IRR, NPV, payback period. 30% ITC + MACRS depreciation modeled. Downloadable Excel proposal report. Built to match SOLON's exact financial modeling requirements.
IRR/NPVEIA APINRELStreamlitExcel
Live Solar / Policy / Market Research HI 38% AZ #5 8% hurdle rate AZ Rank #5 13.0% IRR 7.2yr payback EIA 11.7¢/kWh State IRR ranking · EIA rates + NREL solar · Arizona = SOLON home market
P15 · Solar Incentive Intelligence↗
US Solar Incentive & Market Dashboard
EIA Open Data + NREL Solar Resource + DSIRE
All 50 states ranked for commercial solar economics. Arizona ranks #5 nationally — IRR 13.0%, 7.2yr payback. Real EIA utility rates, NREL solar resource, and DSIRE state incentive programs combined into one composite score.
EIA APINRELDSIREPythonSQL
Live FEMA + Census / Equity $312.7B 2017 Recovery Gap Low poverty 8mo High poverty 39mo 4.7x Annual disaster cost 2000-2023 · 4.7x recovery gap by poverty level
P13 · Disasters / Equity↗
When Disaster Strikes the Poor
FEMA OpenFEMA API + Census Bureau SAIPE
$1.49 trillion in disasters 2000-2023. High-poverty counties take 4.7x longer to recover. Tornadoes hit low-income counties 72% of the time. 2017: Harvey + Irma + Maria = $312.7B.
FEMA OpenFEMACensus SAIPESQLscipy
Live SEC EDGAR / Financials WMT $611B MSFT 34.2% Revenue bars · Dashed = net margin · MSFT 34.2% highest
P11 · Fortune 500 / SEC↗
Fortune 500 by the Numbers
SEC EDGAR XBRL API — Real 10-K Filings
12 companies, 5 sectors, 2019-2023. Walmart leads at $611B but Microsoft wins on margins at 34.2%. Apple generates $100B free cash flow — more than the entire energy sector.
SEC EDGARPythonSQLPOWER()
Live FBI UCR / Investigative STL 65 National avg 6.3 COVID 2020 +56% homicide spike City homicide rates per 100k · National avg 6.3
P12 · Crime / FBI UCR↗
Crime in America
FBI Uniform Crime Reporting Program
St. Louis: 65.4/100k — 10.4x the national average, higher than Brazil and South Africa. COVID caused the largest single-year homicide spike on record: +56% in 2020. Property crime down -21% since 2015.
FBI UCRPythonSQLscipy
All Other Projects
NLP / Healthcare Hospital Review Sentiment Positive 58% Neutral 35% Negative 7% wait staff care billing bias gap MIT License · Python
NLP↗
Hospital Review Bias Detection
Python · MIT License
Detecting service gaps and bias signals in hospital reviews using NLP sentiment analysis.
PythonNLPML
GCP / Analytics 10M+ Ride Records · Looker Studio 10M+ rides GCP pipeline Looker dashboard
Analytics↗
Uber Analytics Dashboard
GCP · Mage.ai · Looker Studio
Full-stack analytics on 10M+ ride records. Urban mobility trends visualized in Looker Studio.
GCPMage.aiLooker
HR / ML Recruitment Funnel Analysis Applications Screening Interview Offer
HR Analytics↗
HR Recruitment Funnel
Python · SQL · Streamlit
Enterprise ML-powered recruitment intelligence platform transforming HR data into actionable insights.
PythonSQLStreamlit
E-commerce / BI KPI Dashboard · YoY Trends $2.4M Revenue +18% YoY Growth 3.2% Conv Rate $84 Avg Order
E-commerce↗
E-commerce Analytics
Python · Plotly · Power BI
Interactive sales dashboard with KPI tracking, YoY trends, and geographic visualizations.
PythonPlotlyPower BI
Jupyter / EDA Coffee Sales by Product Espresso Latte Cappucino Mocha Cold Brew Other
EDA↗
Coffee Sales Dashboard
Jupyter Notebook
Sales analysis and visualization across coffee product categories with trend analysis.
Jupyterpandasmatplotlib
R / Datathon Data Miners Datathon Team project · R language
Competition↗
Data Miners Datathon
R · Team Project
Competitive datathon project with team analysis and data-driven insights using R.
Rtidyverseggplot2
Bioinformatics PSMC Demographic Reconstruction PSMC Analysis ABBA-BABA test Max-entropy dist. Evolutionary analysis Python · Jupyter · Bioinformatics
Capstone↗
Bioinformatics Capstone
Python · Jupyter
PSMC demographic reconstruction, ABBA-BABA testing, and max-entropy distribution modeling.
PythonBioinformaticsJupyter
ML / Classification Spam Email Classifier — Confusion Matrix TP 97.2% FP 2.8% FN TN Naive Bayes + SVM classifier
ML Classification↗
Spam Email Classifier
Python · Jupyter Notebook
Spam classification using Naive Bayes and SVM. Confusion matrix analysis and feature importance.
Pythonscikit-learnNLP
R / Academic Data Science and Public Interest · INFO536 R · tidyverse · ggplot2 · UArizona coursework
Academic↗
Data Science and Public Interest
R · UArizona INFO536
Graduate coursework exploring data science applications in public policy and social interest domains.
Rtidyverseggplot2
R / Group Project Public Interest Research · Group Project Data Public Policy Collaborative R project · UArizona
Team Project↗
Group Data Science Project
R · Data Science and Public Interest
Collaborative research project exploring data science for public interest outcomes.
RCollaborationResearch
R / Research INFO-536 Research Analysis Analysis Insight Impact Policy R · Public interest research · UArizona
Research↗
INFO-536 Data Science Research
R · UArizona Graduate Course
Research on data science applications for public interest and social impact analysis.
RResearchPublic Policy
27
GitHub Repos
13
Projects Built
5+
Live Dashboards
20+
Certifications
Real data.
Real findings.
Real impact.

I'm a Data Analyst and Scientist pursuing my MS in Data Science at the University of Arizona (2024-2025). My work focuses on real government APIs — the same sources that economists, journalists, and policy makers use daily. ROC-AUC 0.978 with XGBoost.

Python
92%
SQL
88%
Power BI
85%
Machine Learning
82%
R
70%
AWS
75%
Best Score
ROC-AUC 0.978
XGBoost model
Location
Tucson, AZ
Open to remote and relocation
Status
Open to Work
Full-time Data Analyst or Data Scientist
Certifications · 20+ Earned
🤖 AI:BI Analyst — Databricks ✨ Gen AI Fundamentals — Databricks 🏆 ML at Scale — Databricks 📊 SQL Analytics — Databricks 🐍 Python Data Science Bootcamp 🗄️ Practical SQL with SQLite 🐍 Python Automation and DS 🧠 Python for AI and ML 🐍 Python Masterclass ☁️ AWS Glue Getting Started ☁️ AWS Data Foundations ☁️ AWS StationOps ☁️ Building LLMs on AWS ☁️ Dynamic Data Masking — Aurora 📊 Power BI Masterclass 📊 Data Analysis and Viz — Copilot 📈 Business Data Analysis 📉 Visualization for Leaders 🐳 Docker Deep Dive 📝 Data Analysis with Copilot
Get In Touch

Let's work
together.

Send an Email GitHub · 26 repos LinkedIn