Publications by Florian Teschner
How to create a ggplot Theme – Unicorn Edition
Themes are an convenient way to give ggplot charts an individualized, sometimes stylish look. Most of the time, I rely on the ggthemes package and the Economist style. Last week colleagues asked me to change the look of my charts. We joked around and I agreed to create a unicorn ggplot theme. I want to use the challenge to detail a) how to create...
3809 sym R (1123 sym/3 pcs) 6 img
Image description with Microsoft’s Cognitive Services and R
A while back, I created the small package Roxford to access Microsoft’s Cognitive Services API in order to easily recognize objects in images. Back then Microsoft called the service “Project Oxford”, hence the name “Roxford”. Since then Microsoft extended their API to include image tagging, description and celebrity detection. In the f...
2358 sym R (331 sym/4 pcs) 4 img
The Instant Rise of Machine Intelligence?
Currently the news are filled with articles about the rise of machine intelligence, artificial intelligence and deep learning. For the average reader it seems that there was this single technical breakthrough that made AI possible. While I strongly believe in the fascinating opportunities around deep learning for image recognition, natural langua...
3004 sym 4 img 1 tbl
From Descriptive to Prescriptive Analytics
Predominately data science projects deal with descriptive statistics. The common theme (especially on this blog) is to gather a dataset, visualize and describe it. The toolset consists of a combination of machine learning, descriptive statistics and (gg-)plots. This time I want to go a step further; from descriptive to prescriptive analytics. Th...
3795 sym R (1291 sym/11 pcs) 2 img
Extending existing packages: Rmisc
One of my favorite packages is Rmisc. The package includes the summarySE function which I use on a daily basis. The function provides a concise way to get a data frame with mean and standard errors of the mean. It is a great way in conjunction with ggplot to visually show differences between groups. Let’s have a look at a play example; we crea...
1945 sym R (942 sym/6 pcs) 4 img
Optimising your blog with R and Google Optimize
When writing a blog, sooner or later, one measures and consequently tries to improve some kind of success metric. In my case; I have been tracking user behavior (visitors, bounce rate, pageviews, …) for slightly over a year with Google Analytics. While some posts have been more “successful” than others, it is not my goal to change the topic...
2173 sym R (990 sym/2 pcs) 4 img
Deep Learning for Brand Logo Detection
A year ago, I used Google’s Vision API to detect brand logos in images. Since then the DIY deep learning possibilities in R have vastly improved. With the release of Keras for R, one of the key deep learning frameworks is now available at your R fingertips. Following up last year’s post, I thought it would be a good exercise to train a “si...
3148 sym R (941 sym/3 pcs) 4 img
How to Scrape Images from Google
In my last post, I tried to train a deep neural net to detect brand logos in images. For that, I downloaded the Flickr27-dataset, containing 270 images of 27 different brands. As this dataset is rather small I turned to google image search and wrote a small R script to download the first 20 images for each search term. In order to use the script...
1165 sym R (89 sym/1 pcs) 2 img
Deep Learning for Brand Logo Detection – part II
A month ago, I started playing with the deep learning framework Keras for R. As a use-case I picked logo detection in images. While the training of a net worked out fine, the results were mediocre. (Check out the full post to for details on the model and the setup.) Here is the recap of the outcome; training the model on the Flickr27-dataset, wi...
2419 sym 6 img
Transfer Learning with Keras in R
In my last posts ([here](http://flovv.github.io/Logo_detection_deep_learning/ and here, I described how one can detect logos in images with R. The first results were promising and achieved a classification accuracy of ~50%. In this post i will detail how to do transfer learning (using a pre-trained network) to further improve the classification a...
3204 sym R (4508 sym/2 pcs) 2 img