HTML + Python / live site
basarorsel.com
HealthFit exports become an embedded training ledger. Python aggregates the record; dependency-free SVG draws volume, routes and inferred routine exposure.
Başar Örsel — portfolio
portfolio · training data · projects · code
The training I logged.
The things I built.
The source behind both.
The person is at basarorsel.me. This is the working record: measured sessions, public projects, and code you can inspect.
01 — Athlete
Returning to competition.
Former national orienteer for Turkey. Rebuilding toward the senior-elite men's team, E21E, in 2027.
Current state: detrained. Previous selection is history, not evidence of current fitness. The work now is restoring running capacity and bringing it back to the map.
The running gate is 3000 m under 10:00. It is a target, not an achieved result or a guarantee of selection.
The log below shows what the watch recorded. It cannot tell you how cleanly I navigated or whether I am ready to race.
Mountaineering adds time on feet and ascent. Triathlon-oriented training currently means swim + run: no cycling data exists in this export. Strength and mobility are load-bearing structures supporting the endurance work, not afterthoughts.
orienteering · mountaineering · swim + run · strength · mobility
E21E · 2027selection target · not achieved
3000 m < 10:00running gate · not achieved
—fastest recorded rolling 5 km · historical
—longest recorded run
Historical bests do not describe today's fitness.
Full training mix / every discipline
Deduplicated sessions, grouped by FIT sport and display category. Unrecognized sports stay visible. Sub-sports retain the export's labels.
| Discipline | FIT sport | Sub-sports | Sessions | Hours | km | Ascent / m |
|---|---|---|---|---|---|---|
| hiking | hiking | generic | 6 | 22.116 | 31.7 | 2,499 |
| jump rope | jump rope | generic | 1 | 0.101 | — | — |
| mind and body | fitness equipment | pilates | 4 | 2.049 | — | — |
| mind and body | training | flexibility training, yoga | 13 | 4.603 | — | — |
| other | generic | generic | 20 | 5.186 | — | — |
| running | running | generic | 129 | 120.599 | 1,010.189 | 3,016 (54/129 recorded) |
| strength | training | strength training | 42 | 14.435 | — | — |
| swimming | swimming | open water | 3 | 1.256 | 1.732 | — |
| walking | walking | generic, indoor walking | 127 | 45.707 | 229.776 | 1,997 (120/127 recorded) |
Hours use recorded timer time. Distance and ascent are sums of recorded fields only; — means unrecorded, not zero. Partial sums show their session coverage. Indoor zero distances are export values, not inferred travel. Display precision is not sensor accuracy.
02 — Training data
The training log.
Volume, pace, heart rate and the gaps between recorded sessions. This is a dated snapshot of watch data, not a live feed.
Read the limits. Rolling splits are GPS-derived, not certified race results. Heart-rate load is a model, not a fitness test. Missing logs are not proof of inactivity. Methods & source ↓
Monthly running volume
Weekly volume
km per ISO week, zero weeks included — gaps stay visible
Aerobic efficiency
quarterly means · pace (min/km, solid) / HR (bpm, dashed) · independent scales; not a controlled test
Recorded runs
runs >500 m and pace <12 min/km · distance (x) / pace (y) · opacity = recency
Fitness · fatigue · form
Aerobic decoupling
speed/HR change between halves · qualifying runs ≥40 min · +5% reference, not a diagnosis
The territory
Outlined break-point markers indicate likely stops / break points: speed below 0.5 m/s for more than 120 seconds. GPS gaps over 30 seconds are excluded; slow movement and GPS error can still look like stops.
Long runs — half marathon and beyond
Records & structure
fastest recorded rolling splits and support work · not race-certified
Mountaineering / hiking
FIT hiking sessions. Recorded ascent, not summit elevation or a claim of technical climbing.
6
sessions
31.7
recorded km
22.116
timer hours
2,499
recorded ascent / m
| Date | km | Ascent / m | Timer / h:mm:ss |
|---|---|---|---|
| 2026-08-04 | 6.248 | 655 | 3:49:18 |
| 2026-08-05 | 6.047 | 1,067 | 6:20:35 |
| 2026-08-05 | 2.349 | 43 | 3:26:57 |
| 2026-08-06 | 5.188 | 350 | 2:57:51 |
| 2026-08-06 | 6.073 | 360 | 3:20:59 |
| 2026-08-08 | 5.796 | 24 | 2:11:18 |
No cycling data exists in this export. Swim + run records do not establish a completed triathlon or a bike split.
Triathlon / swim + run
Swim training alongside the running record; these totals do not establish a completed triathlon.
3
sessions
1.732
recorded km
1.256
timer hours
| Date | km | Timer / h:mm:ss |
|---|---|---|
| 2026-06-15 | 0.969 | 0:36:29 |
| 2026-07-27 | 0.626 | 0:35:38 |
| 2026-08-06 | 0.137 | 0:03:13 |
No cycling data in this FIT export. Swim + run only; no bike leg or bike totals are inferred.
Running record: 129 runs · 1,010.189 km · 120.599 h. Running dashboard →
Strength / muscle heatmap
Inferred by matching workout timestamps/dates against the routine log; a programmatic estimate, not biometric measurement. Historical sessions use the approved routine union, not exercise-level FIT evidence.
Intensity means attributed sessions, not force, fatigue or measured muscle activation. Equal historical intensities reflect the mapping assumption.
Schematic muscle-group map, not anatomical segmentation. Pale = fewer sessions; solid = the largest count in this view. Outlined = no attributed sessions.
Mobility / muscle heatmap
Inferred by matching workout timestamps/dates against the routine log; a programmatic estimate, not biometric measurement. Historical sessions use the approved routine union, not exercise-level FIT evidence.
Intensity means attributed sessions, not force, fatigue or measured muscle activation. Equal historical intensities reflect the mapping assumption.
Schematic muscle-group map, not anatomical segmentation. Pale = fewer sessions; solid = the largest count in this view. Outlined = no attributed sessions.
Methods, limitations & source
Session summaries supply distance and timer time. Duplicate start-time/sport records are removed. Distance totals reconcile against monthly and weekly sums, within rounding tolerance. Weeks without a recorded run remain visible.
Best splits use a rolling distance window with interpolation through record frames. They are not race results. Route drawings use simplified, projected GPS traces; they omit a basemap and are not for navigation. Only location clusters are drawn; other traces are counted separately.
Banister TRIMP uses average heart rate, the highest recorded HR and an explicitly assumed resting HR. The load curves are exponentially weighted averages, initialized at zero. They are not measured fitness or a prescription. The form value is chronic minus acute load.
Decoupling compares speed/HR between halves, after excluding the warm-up and filtering movement records. Quarterly averages combine different routes, conditions and intensities; pace and HR have separate scales. Elevation is watch-derived and can contain altitude error. Calories are watch estimates.
Targets in Athlete are goals, not telemetry. The current-state description is self-reported. This snapshot does not establish readiness for selection.
Read the parser & calculations ↗ · Download this aggregate ↓
02 — Builder
Small systems. Open source.
Shipped sites, hardware studies and public experiments. A repository is evidence of code, not a verified deployment or a claim of sole authorship.
Production
HTML / live site
basarorsel.me
The identity page. One hand-coded document, hash navigation, a local-hour theme and the same signature and visual rhythm as this portfolio.
Hardware / Systems
RFID / prototype archive
Innovation-Registery
An event and society registration experiment with an RFID reader script. Related RHT_RFID code includes Arduino and Python server components. These are prototypes, not verified production attendance systems.
Raspberry Pi / systems study
Small-server experiments
The raspberry-pi-web-server repository records a Pi hosting experiment. RFID and object detection also appear in the InoReg fork below; that implementation has upstream authors and is not presented as my original work.
Experiments / Archives
Python / unvalidated experiment
carbon-footprint
A receipt-analysis and carbon-estimation experiment with Python, a web interface and tests. Its estimates, accuracy and compliance claims have not been independently validated; this is not a certified environmental accounting tool.
Mathematics / ongoing study
Questions before claims
Topology, knot theory, machine learning and LLMs are study interests. No standalone mathematics simulation was verified in the public repository inventory, so none is presented here as a finished project.
Forks / Studying
Fork / upstream credit
InoReg
A fork of NOMANSAEEDSOOMRO/Attendance-System: RFID and object detection using Raspberry Pi. Listed as a study reference, not an original implementation or a verified deployment by me.
Forks / reading other people's code
Agents & embedded systems
AutoGPT and esp-rfid are upstream projects represented by forks in my account. Their functionality and authorship belong to their maintainers; a fork is not a contribution claim.
How this site is built
HealthFit → FIT → Python → JSON → SVG / self-contained HTML.
Raw FIT files stay local. The aggregate contains training summaries and simplified route shapes. No framework, CDN, tracker, account or runtime network request is required.
04 — Identity
Three sites. Different jobs.
Identity
basarorsel.me
The person: mathematics, sport, current direction and contact.
meet the person →
Portfolio / you are here
basarorsel.com
Training data, finished projects and public source code.
see the projects →
Work / Orsel Studio
orselstudio.tech
WooCommerce, AI/search optimization and complete websites. Business belongs there, not in the training log.
talk about work →