Portfolio guide
Find the right project and ask about the evidence behind its results.
Try the chatbot ↗DATA SCIENCE & ANALYTICS PORTFOLIO
Djellon
Pllana.
From complex data
to decisions that make sense.
Exploring real-world questions through Python, SQL, statistics, and machine learning. Every project starts with a question and ends with evidence.
Discover my workINSIDE THE WORK02 / 105
A better forecast.
A measurable difference.
11.5%
lower forecast error than the
previous-week baseline
Bike demand forecastingHistorical holdout · 2012
105Executed projects
16Real-world datasets
105Executed notebooks
Questions first.
Evidence always.
Evidence always.
NEW / AI CHATBOTS
Meet the assistants.
Three practical demos.
One free, open AI model.
Shop support
Explore a policy-grounded support assistant for a fictional online shop.
Try the chatbot ↗Data interview coach
Practice SQL, forecasting, experiments, and model evaluation.
Try the chatbot ↗Powered by Hugging Face Qwen2.5. Runs in your browser after a model download. No API key or account required.
NEW / LARGE CASE STUDIES
Bigger questions.
Deeper investigations.
15.1 million source observations.
Three complete analytical workflows.
Transport operations
NYC mobility operations
12,885,358 source observations
Where are recorded taxi activity and charges concentrated, and can pickup-time information estimate trip duration in a later month?
Hospitality
Hotel cancellation decision system
119,390 source observations
Can a revenue team rank cancellation risk using conservative booking information, and how should it judge a proposed follow-up policy?
Energy planning
Electricity demand planning
2,075,259 source observations
How reliably can a household energy planner forecast the next day while preserving missing meter readings and uncertainty?
01 / SELECTED WORK
Where analysis
becomes insight.
Four studies. Different questions.
The same commitment to evidence.
Commerce · End-to-end
Retail customer analytics
From 541,909 invoice lines to a clear view of sales concentration, customer segments, and repeat purchasing.
61.4%of identified gross sales from the top 10% of buyers
SQL · Pandas · Cohort analysisRead case study
Mobility · End-to-end
Next-day bike demand forecasting
A next-day forecast with calendar-aware features and an honest test on future observations.
11.5%lower MAE than the selected previous-week baseline
Python · Forecasting · Scikit-learnRead case study
Materials · Modeling
Concrete strength prediction
Predicting concrete strength while keeping repeated mix formulations on the same side of the evaluation split.
4.34 MPamean absolute error on unseen mix formulations
Regression · Grouped evaluationRead case study
Manufacturing · Modeling
Multiclass fault recognition
Recognizing steel plate fault types, with target leakage removed and class-level performance made visible.
0.783macro F1 across seven recorded fault classes
Classification · Model diagnosticsRead case study
NEW / POWER BI PROJECTS
Every metric.
Explained and inspectable.
Editable reports, data models, and DAX.
Built around real analytical questions.
Report formats and prepared data are checked. Native Power BI Desktop refresh and rendering remain to be verified. The figures below show the Python analysis.
POWER BI · Commerce
Retail performance with Power BI
Three editable report pages, 17 DAX measures, prepared data, and an executed analysis notebook.
POWER BI · Mobility
Demand reliability with Power BI
Three editable report pages, 17 DAX measures, prepared data, and an executed analysis notebook.
02 / THE PROJECT LIBRARY
105 projects.
More questions. Deeper answers.
Explore by dataset, question, or method.
Every study has a result you can inspect.
Three large case studies, two original end-to-end studies, 98 focused analyses, and two BI extensions across 16 historical public datasets. The NYC case study adds a third native Power BI project. Related studies reuse data to explore different questions.
105 projects
001End-to-end
RetailRetail customer analytics
Where are invoiced sales concentrated, and which customers should be considered for a retention experiment?
Commerce
002End-to-end
Bike sharingNext-day bike demand forecasting
Can historical demand and calendar information improve next-day hourly rental forecasts?
Mobility
003Analysis
RetailProduct credit exposure
Which product codes have the largest credited value relative to invoiced sales?
Commerce
004Analysis
RetailMarket basket associations
Which frequently purchased product pairs co-occur more than their individual popularity suggests?
Commerce
005Analysis
RetailProduct ABC inventory priorities
How many product codes account for the first 80% of invoiced product value?
Commerce
006Analysis
RetailSales data cleaning sensitivity
How do plausible cleaning choices alter gross invoiced sales?
Commerce
007Analysis
RetailObserved reorder intervals
What is the distribution of elapsed time between identified customers’ purchases?
Commerce
008Analysis
RetailOrder size and fulfillment workload
How does product variety per invoice relate to order value and units to handle?
Commerce
009Analysis
RetailWithin-product pricing consistency
Which frequently sold codes exhibit the widest middle-range price differences?
Commerce
010Analysis
RetailNinety-day repeat purchase cohorts
How does repeat purchasing compare when every eligible customer receives 90 days of follow-up?
Commerce
011Analysis
Bike sharingCommuter and casual rider mix
How does the registered-rider share differ by day type and selected hours?
Mobility
012Analysis
Bike sharingWeather and commute-hour demand
How does observed commute-hour demand vary across weather categories and day types?
Mobility
013Analysis
Bike sharingMissing rental-hour coverage
When are calendar hours missing from the rental dataset?
Mobility
014Analysis
Bike sharingDemand persistence over time
At which elapsed-time lags are observed hourly rental counts most strongly correlated?
Mobility
015Analysis
Bike sharingDaily peak concentration
What share of a complete day’s rentals occurs in its four busiest hours?
Mobility
016Analysis
Bike sharingYear-over-year demand comparison
How does observed hourly demand compare between corresponding months of the two years?
Mobility
017Analysis
Bank marketingContact channel response
How do observed subscription rates differ by recorded contact channel?
Marketing
018Analysis
Bank marketingCampaign contact saturation
How does observed subscription vary with the number of campaign contacts?
Marketing
019Analysis
Bank marketingPrevious campaign outcomes
Do prior successful campaign records identify a different response population?
Marketing
020Modeling
Bank marketingPre-call subscription benchmark
Can contact context and prior campaign history predict subscription without call duration?
Marketing
021Modeling
Bank marketingCall duration leakage audit
How much does unavailable call-duration information change retrospective model scores?
Marketing
022Modeling
Bank marketingSubscription probability calibration
Do predicted probability bands match observed subscriptions on later records?
Marketing
023Modeling
Bank marketingCampaign ranking under capacity limits
What precision and recall would observed rankings yield when reviewing a fixed fraction of records?
Marketing
024Analysis
Bank marketingUnknown category coverage
Which fields contain the most explicit unknown values, and does their response mix differ?
Marketing
025Analysis
Bank marketingPrior-contact recency and response
How does response differ between never-contacted clients and prior-contact recency bands?
Marketing
026Analysis
Bank marketingCampaign population drift
How do response and campaign intensity change across source-order quarters?
Marketing
027Analysis
Wine qualityWine quality label balance
How strongly are sensory labels concentrated in the middle of the rating scale?
Food science
028Analysis
Wine qualityAlcohol and quality within wine type
How does median sensory quality vary across declared alcohol bands within red and white samples?
Food science
029Analysis
Wine qualityChemistry associations with quality
Which laboratory measurements have the strongest rank associations with quality within each wine type?
Food science
030Analysis
Wine qualitySugar and density measurement structure
How does density vary across residual-sugar bands within wine type?
Food science
031Analysis
Wine qualityFree and total sulfur consistency
How variable is the free-to-total sulfur measurement ratio?
Food science
032Modeling
Wine qualitySensory quality regression
Can chemical measurements estimate sensory quality better than a median baseline?
Food science
033Modeling
Wine qualityHigher-rated wine screening
Can chemistry rank samples with an observed quality score of at least seven?
Food science
034Analysis
Wine qualityRepeated laboratory signature audit
How often do identical laboratory signatures recur, and can their sensory scores disagree?
Food science
035Modeling
Wine qualityQuality model errors by wine type
Does a pooled quality model have different error or bias for red and white samples?
Food science
036Analysis
Wine qualityUnsupervised chemistry profiles
Do chemistry-based clusters differ in their sensory-score distributions?
Food science
037Analysis
Concrete strengthCuring age and measured strength
How do strength distributions differ across curing-age bands?
Materials
038Analysis
Concrete strengthWater-to-binder ratio analysis
How does observed water-to-binder ratio relate to strength?
Materials
039Analysis
Concrete strengthSupplementary binder usage
How do observed supplementary-binder fractions compare in 28-day samples?
Materials
040Analysis
Concrete strengthSuperplasticizer and water demand
How does water content differ across observed superplasticizer dosage bands?
Materials
041Modeling
Concrete strengthConcrete strength prediction
Can mix measurements predict strength beyond a median baseline on unseen formulations?
Materials
042Modeling
Concrete strengthStrength prediction by curing age
At which curing ages are the fixed strength model errors largest?
Materials
043Analysis
Concrete strengthMix measurement redundancy
Which input variables carry overlapping rank information?
Materials
044Modeling
Concrete strengthLow-strength screening benchmark
Can mix and age measurements rank specimens below a declared analytical strength threshold?
Materials
045Analysis
Appliance energyHousehold appliance load profile
Which hours have the highest typical appliance energy per recorded interval?
Energy
046Analysis
Appliance energyWeekday and weekend load
How do interval energy distributions differ across weekdays?
Energy
047Analysis
Appliance energyDaily household energy totals
How variable is daily appliance use when only complete days are counted?
Energy
048Analysis
Appliance energyEnergy concentration in high-load intervals
What fraction of appliance energy occurs in the highest-load recorded intervals?
Energy
049Analysis
Appliance energyIndoor conditions and appliance use
Which recorded indoor conditions have the strongest rank associations with appliance energy?
Energy
050Modeling
Appliance energyOne-hour-ahead appliance forecast
Can lagged energy and known calendar features estimate appliance use an hour ahead?
Energy
051Analysis
Appliance energyRandom-variable negative controls
How do source-supplied random controls compare with measured variables in an association screen?
Energy
052Modeling
Appliance energyHigh-load interval warning benchmark
Can past appliance use rank unusually high-load intervals one hour ahead?
Energy
053Analysis
Air qualityAir monitoring data availability
Which sensor and reference fields have the weakest observation coverage?
Environment
054Analysis
Air qualityMonitoring outage duration
How long are the longest consecutive missing runs in each monitoring channel?
Environment
055Analysis
Air qualityHourly carbon monoxide pattern
What daily pattern appears in observed reference carbon monoxide measurements?
Environment
056Analysis
Air qualityMultisensor cross-sensitivity
Which sensor responses are most associated with each available gas reference?
Environment
057Modeling
Air qualityCarbon monoxide sensor calibration
Can contemporaneous sensor responses estimate an available reference CO concentration?
Environment
058Analysis
Air qualitySensor association drift
How stable is the relationship between a CO-sensitive sensor and the reference over calendar months?
Environment
059Modeling
Air qualityOne-hour-ahead CO forecast
Can earlier observed CO and sensor responses improve a one-hour-ahead reference forecast?
Environment
060Modeling
Air qualityHumidity and sensor calibration error
Does the fixed CO calibration model perform differently across relative-humidity bands?
Environment
061Analysis
Online shoppersSeasonal session conversion
How does observed purchase conversion vary across recorded months?
Commerce
062Analysis
Online shoppersReturning and new visitor behavior
Do recorded visitor types differ in purchase conversion and product engagement?
Commerce
063Analysis
Online shoppersTraffic source quality
Which recorded traffic categories combine meaningful volume with higher observed conversion?
Commerce
064Analysis
Online shoppersProduct browsing depth
How does completed-session product browsing depth relate to conversion?
Commerce
065Analysis
Online shoppersBounce and exit measurement audit
How do bounce and exit rates relate, and where are their recorded values unusual?
Commerce
066Modeling
Online shoppersCompleted-session purchase classification
How well can a model classify purchase sessions without PageValues?
Commerce
067Modeling
Online shoppersPage value information audit
How does PageValues alter retrospective purchase classification?
Commerce
068Analysis
Online shoppersBrowser conversion diagnostics
Which browser categories warrant investigation of their purchase experience?
Commerce
069Analysis
Fuel economyHistorical fuel economy trend
How does observed city-cycle fuel economy vary by model year?
Mobility
070Analysis
Fuel economyVehicle weight and fuel economy
How does fuel economy differ across declared vehicle-weight bands?
Mobility
071Analysis
Fuel economyEngine configuration comparison
How do cylinder groups differ in observed fuel economy and vehicle weight?
Mobility
072Analysis
Fuel economyOrigin and model-year composition
Does the origin comparison change when model years are grouped into comparable periods?
Mobility
073Modeling
Fuel economyFuel economy prediction
Can technical specifications estimate historical city-cycle fuel economy?
Mobility
074Modeling
Fuel economyLater-model fuel economy transfer
How do specification-based models perform on later model years?
Mobility
075Analysis
Fuel economyHorsepower missingness audit
Which model-year and origin groups contain missing horsepower measurements?
Mobility
076Analysis
Fuel economyFuel consumption unit conversion
How do MPG summaries compare with fuel used per 100 kilometres?
Mobility
077Analysis
Wholesale customersWholesale channel assortment
How does the median spending mix differ between recorded sales channels?
Commerce
078Analysis
Wholesale customersCustomer spending concentration
What fraction of aggregate annual spending is associated with the highest-spending customers?
Commerce
079Analysis
Wholesale customersWholesale customer segmentation
Do customers form interpretable groups based on annual spending patterns?
Commerce
080Analysis
Wholesale customersCross-category spending relationships
Which spending categories show the strongest customer-level rank associations?
Commerce
081Analysis
Wholesale customersRegional customer basket comparison
How does observed total annual spending differ across recorded regions?
Commerce
082Analysis
Wholesale customersAtypical wholesale purchasing profiles
Which customer spending profiles are unusual relative to this sample?
Commerce
083Analysis
Steel plate faultsRecorded fault class balance
Which defect classes dominate the labeled inspection sample?
Manufacturing
084Analysis
Steel plate faultsFault geometry profiles
How do pixel area and perimeter profiles differ by recorded fault class?
Manufacturing
085Analysis
Steel plate faultsThickness and fault composition
How does the composition of recorded faults vary across plate-thickness bands?
Manufacturing
086Analysis
Steel plate faultsInspection luminosity profiles
Which fault classes have different observed image contrast and luminosity profiles?
Manufacturing
087Modeling
Steel plate faultsMulticlass fault recognition
Can inspection features distinguish the seven recorded fault classes?
Manufacturing
088Modeling
Steel plate faultsFault recognition confusion analysis
Which true fault classes are most often confused with another class?
Manufacturing
089Modeling
Steel plate faultsCompact inspection feature set
How much recognition performance changes when only compact geometric features are retained?
Manufacturing
090Analysis
Steel plate faultsInspection geometry integrity
Do image measurements contain impossible bounds or nonpositive areas?
Manufacturing
091Analysis
Forest firesRecorded fire seasonality
Which months contain the most fire records and recorded burned area?
Environment
092Analysis
Forest firesBurned-area concentration
How much of the recorded burned area comes from the largest observations?
Environment
093Analysis
Forest firesZero recorded area analysis
How does the share of zero recorded area vary by month?
Environment
094Analysis
Forest firesWeather and burned-area associations
Which recorded weather and fire-index variables associate with burned area?
Environment
095Analysis
Forest firesSpatial concentration of recorded area
Which source grid cells contain the greatest recorded burned area?
Environment
096Modeling
Forest firesBurned-area severity benchmark
Can recorded conditions estimate burned area beyond a simple baseline?
Environment
097Modeling
AbaloneAbalone ring-count estimation
Can physical measurements estimate observed ring counts beyond a median baseline?
Biological data
098Modeling
AbaloneNon-destructive measurement tradeoff
How does restricting predictors to external measurements change ring-count error?
Biological data
099Analysis
AbaloneBiological measurement plausibility
Which observations violate simple dimensional or mass consistency checks?
Biological data
100Analysis
AbaloneShell shape and specimen groups
How do shell-shape ratios and observed rings differ among recorded specimen categories?
Biological data
101Power BI
RetailRetail performance with Power BI
How can a commercial team monitor sales, credits, and customer coverage without ambiguous KPI definitions?
Commerce
102Power BI
Bike sharingDemand reliability with Power BI
Where does a better next-day demand forecast still underpredict, and how much of the evaluation calendar is observed?
Mobility
103Capstone
NYC taxi tripsNYC mobility operations
Where are recorded taxi activity and charges concentrated, and can pickup-time information estimate trip duration in a later month?
Transport operations
104Capstone
Hotel bookingsHotel cancellation decision system
Can a revenue team rank cancellation risk using conservative booking information, and how should it judge a proposed follow-up policy?
Hospitality
105Capstone
Household electricityElectricity demand planning
How reliably can a household energy planner forecast the next day while preserving missing meter readings and uncertainty?
Energy planning
No projects match those filters.
Try a different keyword or dataset.
03 / MY APPROACH
Curiosity, with
a rigorous side.
I use data science to connect a real question with an answer that can be checked. My work spans customer behavior, forecasting, quality analysis, and model evaluation—with the assumptions in view.
01
Start with the question.
Define the decision, the available information, and what a useful answer would change.
02
Make the checks count.
Trace sources, preserve missingness, prevent leakage, and compare with a simple baseline.
03
Show the whole result.
Share code and findings alongside uncertainty, limitations, and the next test worth running.
PythonSQLPower BIDAXPandasScikit-learnStatisticsTime seriesData visualization
BUILT TO BE EXPLORED
Go beyond the chart.
Browse the source, open an executed notebook, or reproduce an analysis.
The complete collection includes setup instructions and dataset attribution.