I'm Alexander, a biologist turned Java Software Developer with a passion for software that solves problems and automates stuff.
After completing my master's degree in Biology, I found my way into the world of IT, where I've been thriving ever since. Here, I discovered that the systematic approach from biology is really useful for technology too. The world is a gigantic system afterall!
Here, you'll find a collection of programming projects I've worked on, ranging from Java applications, over scripts and data science projects to web development ventures. Feel free to explore the code on GitHub or dive into the summaries provided below.
- 🔭 I’m currently working as a Java Developer for a Document Management System.
- 🌱 I’m passionate about agility methodologies and the evolving landscape of Web3 technologies.
- ⚡ Fun fact: When I'm not coding, you'll find me immersed in fantasy novels (or trying to write one myself), strumming away on my guitar, or pondering the intricacies of philosophy.
- 📫 How to reach me: Linkedin Gmail Twitter
- 👨💻
- 🌐
- 🛠️
- 🛢
- 📚
- 🏗️
- 🎨
Show projects
This is my Issue Tracker. It is still developing. Right now, it has a List of Users and a Navigation-Bar.
More details
The goal is to have a completely running Issue Tracker comparable to GitHub running.The core and planned features are:
| Core | Advanced |
|---|---|
| Creating an issue | User authentication |
| Viewing issues | Assigning issues |
| Updating an issue | Sorting issues |
| Deleting an issue | Filtering issues |
| Pagination | |
| Dashboard |
This is my fullstack application with Java and Angular. It is split up into frontend and backend:
More details
I created this application as a possible starting point and as a reference for new projects. It`s running completely and works with MySQL.
My Todo List application with Javascript, node.js, and express.js.
More details
This is a Todo List application built with Node.js, Express, and MongoDB. It allows users to create, view, and delete tasks.
This is my Weather application.
More details
Get your own API-Key (it's for free), run the application (1. node install, 2. node ./app.js), and you can check the weather on your own website!
This was my first complete Web Design course.
More details
I got my FreeCodeCamp Responsive Web Design Certificate!For this I went through the course (estimated 300 hours) and finished these 5 projects.
Here, I updated my trial website and made a CatPhotoApp with HTML and CSS.
This is my first HTML Website. I created a simple pancake recipe and experimented with some HTML code.
My ANOVA analysis with python of a freely available tips-dataset.
More details
* Explored the data * Boxplots * QQ-plots * Histograms * Used different R libraries for data exploration * Controlled the data for outliers * Compared if the total bills are statistically significantly different per weekday.H1: The weekday has an effect on the amount of the total bill. (M[1] != M[2] != M[3] != M[4])
H0: The weekday has no effect on the amount of the total bill. (M[1] = M[2] = M[3] = M[4])
The weekday has a significant influence on the total bill (F(3, 240) = 2.767 , p = 0.0424). 3.34 % of the spread of the total sum can be explained by the weekday. According to Cohen (1988) is the effect size of 0.186 a small effect.
The Bonferroni Post-Hoc-Test shows that no groups can be generalized out of the weekday (all p > 0.05). Thursday (M=17.68, SD=7.89, N=62), Friday (M=17.15, SD=8.30, N=19), Saturday (M=20.44, SD=9.48, N=87) and Sunday (M=21.41, SD=8.83, N=76) are not significantly different.
It can be concluded that no independent groups can be formed that differ from each other. Hence, although the ANOVA was significant, the H0 is kept and H1 declined.
My statistical exploration with python of a freely available tips-dataset.
More details
* Explored the data * Boxplots * QQ-plots * Histograms * Used different Python libraries for data exploration * Controlled the data for outliers * Used Pearson to check if there is a correlation between the total bill and the tip * Compared if man and woman give significantly different tips.The tip and the total bill correlate positively significant (r = 0.6757341, p < 2.2e-16, n = 244). Hence, it can be said the higher the total bill the higher the tip. 45.66 % of the spread of the whole variance can be explained through the tip and the total bill. According to Cohen (1992) is the effect size of 0.68 a strong effect. The H0 can be discarded.
There is no significant difference between the tip of women (M = 2.83, SD = 1.16, n = 87) and man (M = 3.09, SD = 1.49, n = 157), (t(242) = -1.3879, p= 0.1665, n=244)). According to Cohen (1992) is the effect size of 0.185 no effect. The H0 cannot be discarded.
My statistical exploration with python of a freely available freedom-status-dataset.
More details
* Explored the data * Boxplots * QQ-plots * Histograms * Used different R libraries for data exploration * Controlled the data for outliers * Compared if the different Freedom status are statistically significantly different groups.H1: There is a mean difference between degree of freedom and the relative GDP. (FS[F] != FS[NF] != FS[PF])
H0: There is no mean difference between degree of freedom and the relative GDP. (FS[F] = FS[NF] = FS[PF])
The degree of freedom has a signifikant influence on the relative GDP (F(2, 49.062) = 15.491 , p = 6.097e-06). 19.4 % of the variation in the relative GDP around the overall mean can be explained by the degree of freedom. The effect strength according to Cohen (1988) is f = 0.4907 and corresponds to a strong effect.
The Tukey post-hoc-test shows, that two groups can be constructed according to degree of freedom (all p < .05): free (M = 27841401 , SD = 26632852, N = 56), and not-free (M = 12368148, SD = 18508685,N = 21) and partially-free (M = 6202767 , SD = 10345156, N=44) form the second group.
It can be concluded that two independent groups can be formed that differ from each other. Freedom is the most effective. H0 is declined, H1 accepted.
More details
* Imported 20000 pictures * Cleaned and reshaped the images * Explored/ visualized the images * Performed different prediction models: * Linear regression * Decision tree * Random Forest * Simple Neuronal Network * Try of convoluted Neuronal Network * Tried different evaluation methods * Heatmap with errors * Loss over epochs * Accuracy
In this project, I compared the legendary pokemon in R.
More details
* Explored the data * Controlled the data * Compared if legendary pokemon have statistically higher attack than non-legendary pokemon.H0: The attack of legendary pokemon is lower or equal than of non-legendary pokemon. (M[L] </= M[NL])
H1: The attack of legendary pokemon higher than of non-legendary pokemon. (M[L] > M[NL])
Legendary (M = 116.68, SD = 30.35, n = 65) pokemon have significantly higher attacks than non-legendary (M = 75.67, SD = 30.49, n = 735) pokemon (t(798) = -10.397, p = 2.513e-05, n= 798). Hence, the H0 can be discarded. The effect-size is d = 3.438535 and is according to Cohen (1988) a very strong effect.
In this project, I used machine learning for predicting asteroid diameter.
More details
* Cleaned the data * Explored the data * Performed different prediction models: * Linear regression * Polynomial regressions * Decision tree * Random Forest
Here, I created different regressions and compared them.
More details
* Used pipelines * Used convenience functions for calculations * performed: * Linear regression * Quadratic regression * Cubic regression * Made different plots
Linear regressions are cool.
More details
* performed: * linear regression * multiple linear regression * polynomial regression * exploratory Data analysis * Made different plots
More details
- I constructed a database from scratch
- I created all necessary tables
- I made all the necessary connections and indices
- I created views, procedures and triggers
- I made useful table and scalar functions
More details
- I scraped data from a website
- I explored the data
- I included everything in objects (OOP)
- I retained graphs from my classes
- Sub plots
- Regression line
- I made my first small prediction
More details
* I explored data from an online database * I extracted different dataframes * I merged important dataframes * I made * Bar charts * Pie chart * Box plot * Sub plots * Different sub plots in one figure
More details
* I made a calculator that calculates the compound interest and plots the graph from start till finish * I took the user input * I made functions to * retain the CI * save the x-values * calculate the compound interest * I made lists for x- and y-values * Finally, I plotted the graph with matplotlib
This was my very first project.