Hi! My name is Alihan Büyükeken. I am a Civil Engineer who loves project management and data analysis. I have been working as a Technical Consultant in Penguen Consultancy. My main job is configuring document control and project management systems for project requirements and training the project participants. I also assist the project management team in the determination, planning, storage, and analysis of data to be collected from the field during the project phases.
I am planning to use my future data science skills to create more data-centric systems and using power of data for project management. Also I do data visualizations in amateur way like a hobby, being pro in visualization can be a plus for me.
Check my LinkedIn Profile
Manage your Shiny app users by Alexander Matrunich
In this session Alexander shows us an useful approach to user management of Shiny apps. Managing users in Shiny can be problematic when we use free version of Shiny Server. Alexander’s solution to this problem is using user management service of a tool called Backendless. We can register new users, authenticate users and track their usings. He talks about their package “backendlessR” that using the tool’s API calls and how we can use that package. I like this session because I have faced similar user management problem like 3 months ago and my solution wasn’t Shiny enough as his solution.
This post is about how to use R for big data. It shows three ways of working big data in R and compares these methods with advantages and disadvantages. First method downsampling gives us speed and prototyping ability but has possible sampling and scaling problems. Second method splitting data to the chunks and analyzing them chunk by chunk is good with parallelization ability but pulling all the data can be time-consuming. Last method is compressing data in the database then pulling to R gains plus point with using database for filtering but this method depends too much on the database type which we use. I learned that there is no greater method which works properly in every situation, there are some methods with pluses and minuses. We need to look at what we have and choose the optimum method regarding this.
You can read this post here
First of all, I like tree-based models, I am very fond of them. This post starts with explaining what is a rule and shows some examples in R. After that, we learn about Ross Quinlan’s C4.5 algorithm and its successor C5.0. C5.0 has more classification rules and inclusion of boosting. Then we learn about another methodology developed by Quinlan which is called Cubist. The main difference with Cubist and C5.0 is Cubist predicts numbers but C4.5/C5.0 predicts categories.
You can read this post here
Another Shiny content because there is never enough praisement for Shiny. This post gives some cool hints for structuring Shiny apps. You can create files of manageable parts from your code and work in a more lean way. You can also learn about using .js in Shiny with this post.
You can read the post by clicking here