Some of my IT-projects

3 GPS programs made by combining js and php

*Mark-up a route on GoogleMap, then distances and elevations for the route points are calculated, and you can generate a gpx file for the route

*Create a gpx file for a OSM(OpenStreetMap) path or from a (lat,lng)-set, elevation profile is also calculated

*Show map witn colored height levels, user interactive scales

gpx is Garmin’s geographical file format, you can download the gpx file to a phone or smartwatch.
You can easily convert a gpx file to i.e. a kml file, which can be opened in GoogleEarth.

Programme to show graphs for tcx and gpx -exercise file

I have developed an HTML file with JS-code, Exercise-Graphs-application.htm, where you can upload a TCX or GPX exercise file to generate detailed graphs for velocity, pulse, and altitude data, for both cycling and running.

Built with JS in collaboration with Gemini AI. After in-depth analysis, I have implemented GAP-velocity (Grade Adjusted Pace) corrections for uphill/downhill segments—for instance, running at 5% incline at 12 km/h is equivalent to 13 km/h on flat ground.

Since altitude data from my Garmin Forerunner 255 (barometric) features slight 0.2m variations per second, I have implemented a weighted moving average algorithm to ensure smooth and accurate elevation and velocity curves.

The app allows you to zoom into specific time intervals to analyze performance and retrieve average values. You are welcome to download and save the HTML file for your own use.

Applications using OpenLayers

https://openlayers.org/en/latest/examples/ shows more than 200 examples, based on javascript using the open source OpenLayer package. It can bet used without installation of gis programs on user pc. Impressing possibilities! incredible number of different data sources can be used. It is easy to make user interactive gis programs. I have tried, tweaked, investigated most of these examples.
Some of the more useful, interesting, impressing examples are shown in the link

BI & Databehandling med Microsoft PowerBI; data-cleansing with Python

Jeg har implementeret min egen Multidimensionale, Data-cube, model

Arbejde med Excel og Power Query

AI og Machine-learning

n&n dataflow tool, ChatGPT, developement of useragents with openai Python-SDK-package, ML-algorithms

In the last weeks I have worked with a rather impressing data-flow-programme n&n.
It facilitates construction of data flows by composing nodes on a visual platforme, thus making it easier to automate data flows with inclusion of AI and to visualize flows to colleagues.

In May, I spent three weeks using ChatGPT for code generation and developing user agents, using OpenAI’s Python SDK from spring 2025. According to experienced developers, this SDK makes many development tasks significantly easier than before —and I agree that many tasks now feel rather more accessible.´ Based on web searches, helpful YouTube videos, and many good examples from OpenAI, I have implemented an orchestrator agent that delegates tasks to other user agents, GuardRails to check user input/output, and tested the latest, highly accessible OpenAI RAG tool to create a user agent that incorporates data from a file directory. I have also experimented with agents that automatically perform web searches —quite impressive.

3 short online courses about Machine-learning(ML)/AI, offered by Stanford University< Supplied by a course in Copilot-code-generation, where I worked with implementation of different ML algorithms in Python: I have a good theoretical and practical understanding of neurale network, super/unsuper-vised ML, linear models, cluster determination, anomaly detection and more.

Contact me: bjarnephys@yahoo.dk