Gokce EZEROĞLU,sales analyst at Avon Cosmetics. I have been working at business analysis department since 2017. My motivation for learning data science is to have broader understanding on deep dive analyses. Currently, programs like python, R, sql, are already being used in my team at Avon. I will be contributing to programming based analyses at the end of this program.
As a part of the first assignment, I watched a video, named ’Automated Visualisations for Big Data.(https://www.youtube.com/watch?v=ZudrjJz8OA0) It is about how to handle, visualize a large data set for coral reefs. R is used for that aim. MysQL, which has an easy integration with R, is used to store data. To avoid from long PDF documents or html(which might necessitate continious scrolling) dynamic plotting is used to explain results. In the video,researhers also used LEAFLET, which allows users to see where exactly the survey take place on map. The most interesting part in this video is “parameteresing R MaRKDOWN documents and RMySQL”It allows users to run a certain part of R code, with the specified parameters. Final thing,explained in the video is, “child documents”.Child documents can be used like text files. When the code is run for the next time, child document will automaticly be reflected to output.
KPI reports, dashboards, visualization is important for an analytics team. I found out ‘Shiny’ which is the web application for R. (https://www.youtube.com/watch?v=1MHX1s5bb6w).Shiny allows to make either static or dymamic dashboards.In my current position, sometimes it is important to show real time data, such as billed orders based on region, sales figures recorded at the system etc.In this video, creating a good looking dahsboard from scratch is well explained. The minimum requirements to start doing dashboard is to have shiny and shinydashboard libraries.
Forecasting is one of the daily routines. You need to understand the trend, and eliminate the seasonality& holiday impact.There are several packages in R, which are used for forecasting. I watched a video, which explains ARIMA modelling.(https://www.youtube.com/watch?v=a5A0dTg5T4k).Arima stands for Auto-Regressive Integrated Movig Average.Later,I wanted to have broader understanding on Arima, and found out pros and cons of it.(http://www.ecostat.unical.it/tarsitano/didattica/SeStoCor/SeStor%2027.3/08notes5GOOD.pdf) In the provided link, it is stated that Arima does not have automatic update functionality. However, it does not have a negative impact on my side, ı find it very useful.
I have watched video about reporting on R markdown. (https://campus.datacamp.com/courses/reporting-with-r-markdown/chapter-one-authoring-r-markdown-reports?ex=1) There are several formats that you can publish your data. You can also add Shiny elements to Rmarkdown file to create interactive presentations. This is the thing I like the most. I also came accross with flexidashboards in data camp. (https://campus.datacamp.com/courses/building-dashboards-with-flexdashboard/dashboard-layouts?ex=3). I found it very useful as well. It looks user friendly and easy to create. Since company data is confidential,I also looked for an answer to put a password into R script. I found the answer in the following page.(https://www.rdocumentation.org/packages/getPass/versions/0.2-2/topics/getPass)