I just had to quick blog on these key take aways from Freelance Analytics Meetup. For freelance gigs visit alpha-zetta.net Great books! Small Data by Martin Lindstrom Data'ism Putting your data to work Actionable Analytics Summary: Regression Analysis - for measuring magnitude and trend Logistic Regression - measuring probability or likelihood Clustering - similarities and groupings Survival Analysis - determining the time to an event Event Analysis - understanding sequences of complex events Network Analysis - determining the connections between entities Kahoot.

While there is the determined attempt of the technical world - experts,consultants, technical advisers, academic institutions in creating data science courses, offering data science services to monetize and capitalize on the current data science “hype”, reading a blog like Fronkonstin created by Antonio Sánchez Chinchón of Telefonica - featured in a Data Camp course on phyllotaxiz, for someone like me who thinks and breathes R, it’s just a ‘breath of fresh air’.

It has often bewildered a great many RF Engineers when it comes to answering this question, “How do you measure the cell throughput during the busy hour?”. There are two camps, the first (1) would go about by getting the sum of all data transferred per QCI (in Mb) and divide it with the sum of the time for each data transfer per QCI (seconds) duration, the other (2) would get the individual transfer rates per QCI and the sum of all transfer rates would be the cell’s data transfer rate at any given hour - which I think is the correct one considering the ridiculous results you get when using the first method.

Leaving an R Trail is what I am set to do. Having done my first flexdashboard in RMarkdown and looking at that gorgeous, modern-feel, professional, expensive looking report not to mention the availability of R under the hood - I am fully persuaded that I will be an R person from here on - and an R Studio fan!… they just make the coolest stuff. My first R script was created last November, that is 107 days up to the day of this post and I must say that my interest for doing everything in R has not waned.

It’s most interesting when you put RQI's (Radio Quality Indicators) side by side and see how they influence your throughput. The following table serves to illustrate the range of throughput fluctuations per bandwidth at varying radio conditions: At extremely low CQI ranges, equivalent to poor radio interface quality, the fluctuations are rather large (10%-30%) representing a very bad perceived experience in terms of throughput. It is therefore suggested that for CQI targets should be set around CQI=10 to guarantee excellent subscriber experience.