RSSAmplifier

Blog

Tomaž Weiss

Data science with R.

tomazweiss.github.ioRSS feed ↗10 posts

Latest posts

The Decline of Stack Overflow

For many years, Stack Overflow has been the go-to platform for developers seeking solutions to coding problems. However, with the rise of AI-driven coding assistants, the platform’s relevance has taken a hit. In this post, we’ll explore the data behind Stack Overflow’s decline, comparing trends in question volume and shifts in popular tags after the introduction of ChatGPT and other LLMs. Other…

Adding ROracle Package to the rocker/tidyverse Docker Image

ROracle is an open source R package supporting a DBI-compliant Oracle driver based on the high performance OCI library. It enables highly scalable and performant connectivity to Oracle Database for data transfer. The installation of this package, however, can be a bit challenging. This post provides a Dockerfile, based on the rocker/tidyverse image, that adds ROracle package. A similar script can…

2023 Developer Compensation by Country

Exploring data from the Stack Overflow Developer Survey 2023 with simplified boxplots. We should be aware that this survey data is biased. Countries with less than 30 responses after applying all filters have been removed from the image. Found this content interesting or useful? Fuel future articles with a coffee! The Code # renv::init() library(tidyverse) library(janitor) library(countrycode) #…

Generating Slovenian Names with Keras/TensorFlow

Generative Deep Learning Recurrent neural networks can be used to generate sequence data such as text. The usual way to generate sequence data in deep learning is to train a network that predicts the next token in a sequence using the previous tokens as input. Let’s explore how this method can be used to generate new Slovenian–sounding names. The same approach could be used for generating original…

Oglaševane prodajne cene stanovanj skozi čas

Višanje cen nepremičnin je zadnje čase aktualna tema, zato si poglejmo, kako se to odraža v nepremičninskih oglasih. Podatki Podatki so bili pridobljeni med marcem 2019 in majem 2021. Tako kot v prejšnjem prispevku, je bila tudi tukaj uporabljena avtomatizirana skripta, ki je enkrat na dan s portala nepremicnine.net pobirala nove oglase. Poudarimo pomanjkljivosti, ki so v tako pridobljenih…

Croatia Earthquake on Twitter

There was a strong earthquake on 29. 12. 2020 in Croatia that was felt all over the region. Let’s see how people tweeted about it. Download the Tweets library(tidyverse) library(reshape2) library(rtweet) library(scales) library(lubridate) We will download the tweets that include the word “earthquake” in Slovenian, Croatian, Serbian, Bosnian, English, German, Italian or Hungarian language. I have…

Object Detection with R on Windows

Package image.darknet Computer vision is becoming more and more accessible. One simple way to try it out by yourself has been made possible by the R package image.darknet which provides an easy to use interface to image classification and object detection based on darknet. A blogpost by AbdulMajedRaja RS explains how you can use this package with a Tiny YOLO model for object detection on images.…

Oglaševane cene stanovanj

Podatki Prodaja Ljubljana in Maribor glede na velikost stanovanja Ljubljana po upravnih enotah Oddaja Ljubljana in Maribor glede na velikost stanovanja Ljubljana po upravnih enotah Donos iz oddajanja Cene nepremičnin so v zadnjem času v Sloveniji zelo aktualna tema. Poglejmo si, kaj o njih pravijo podatki z našega največjega nepremičninskega portala. Podatki Podatki so bili pridobljeni med marcem…

Weather in Ljubljana in 2019

On a hot summer day, I re-created this data visualization: The original with R source code can be found here. Historical weather data for Slovenia is available here. Found this content interesting or useful? Fuel future articles with a coffee!

R Users’ Salaries from the 2019 Stackoverflow Survey

Introduction Stackoverflow has recently released the anonymized results of their 2019 annual developer survey. Let’s take a look on what R users are saying about their salaries. Note that the following results could be biased because of unrepresentative and in some cases small samples. Data Preparation The dataset consists of survey answers from nearly 90,000 respondents. About 5,000 of them…