The need for data analysis has grown so much over the years. I started my data analysis career running away from managing debits and credits as an assistant financial officer. It was over time that the whole world realized that the isn’t enough attention given to analyse all data we have. To add to the mix, smartphones and other portable devices made doing everything easier to the point that we all started using them. In so doing, we now started to collect even more data. The need for a better and faster way for data analysis started being a career.
Everyone who started with data analysis and visualization back in the 2010s remembers being pulled in out of a need. There wasn’t formal training needed for one to be an analyst. All that was important at the time was selecting diligent workers who are good at operating data manipulation applications and software. We all started being given different titles from business analyst, data analyst and in my care, sales operations reporting manager. The word manager in my title attracted two subordinates. It mattered not what our titles were, we were all somehow managing data at different levels.
I recently went online and realized that the data field has attracted many different people. Everyone is asking for the magic formula that is needed for one to be an analyst of data. Some advice is there for the views because I know for a fact that they won’t make one an analyst. Neither are the steps I took as at the time there was a need and everyone even with mediocre analytical skills could have a good career. But today things are different. Universities have been called in to assist with formal training thus the never-ending list of qualifications with data and analytics variations in the title. I did one in big data analytics; I have never needed to analyse big data.
The big data courses are the ones I am particularly interested in. This is when I bumped into a five-page document created to assist anyone who wants to be a data scientist. The document is well-written in big fonts. This ensures that everyone can follow the process. Although I soon realised that what was written on this document is like what is said on social media. However, there is a need for this type of document as we are not all in the same social media circles. Ten steps are typically highlighted for one to follow if you are thinking of a data science career.
Math Fundamentals
It is always encouraged that each school leaving learner should have math basics. I know we are all different and aspire for different things given the complexities we are faced with – hence why others opt out of math and science subjects in South Africa. If you are one of those who opted otherwise, don’t be distressed. Various resources could assist you with understanding concepts of linear algebra, calculus, optimisation, and functions. Don’t worry about having a doctorate equivalent in any of the concepts. Remember that the software will do the work. You just must learn the concept.
Learn Programming
Learning math concepts helps me to translate them into computer language. This is the equivalent of using a calculator instead of scribing the entire math solution on paper. Calculators are handy especially when I don’t have scribbling tools or can’t keep up the sum in my head. Learning how to program is equivalent to me learning how to compute complex math problems using a calculator. Knowing how the computer needs to be punched (inputted) improved my output tenfold. I decided to enrol on an IT management degree at the university and I learned VBA and SQL programming languages. I then took big data in my post-graduate degree. There are options for self-study now or even guided online tutorials to assist with pace. Others suggest online that Python and CS fundamentals could yield better results.
Data Wrangling and Visualisation
Learning how to code assists with data wrangling. Data wrangling techniques include data collection, cleaning the data, and data exploration. This step is important in case you are appointed to the project as a data scientist and there are no data engineers or data analysts to assist with data wrangling. Data visualisation will also assist in data storytelling. You must remember that data storytelling is about what happened to that data along the path to showing results. This will include how data was received, cleaned, and truncated to remove duplicates and all other steps taken before analysis could commence. Being able to visualise this would help keep the audience engaged.
Statistics and Probability
Another way to keep an audience engaged is by showing statistics and probabilities. The best way is to learn how to conduct descriptive statistics. I do descriptive statistics just to let my audience know the min, max, mean, and mode. There are other statistical concepts that we learned in the first year of my finance degree at university that could be extremely useful. This includes inferential, associative, and differential. Probability is extremely important as I learned the hard way when I was writing my master’s dissertation (thesis). Conditional probability and Bayes Theorem could do wonders too. I was lucky my supervisor was there to guide me to use the correct statistical algorithms using SPSS and SAS. At least I knew much about data and storage although I was new to these software tools.
Conclusion
Learning math at school is daunting. To add to the mix, it doesn’t mean that all those with math when matriculated have a monopoly on data analysis or data science careers. We are all given a fresh start anytime we need one. The same is true for all others who didn’t get a chance to sit by the computer in their youth. Others might have had an advantage earlier in life, but we can all learn to code now. It’s even easier now than ever to have access to coding tools and training. Learning stats is extremely beneficial nowadays. I love knowing about ways to get an answer quickly. That is when I realised that stats could help me keep my audience engaged throughout a boring data talk. There are frameworks we should aim to know when it comes to data. Allow me to categorise data wrangling and data visualisation as another framework that is needed for us to be successful data analysts/scientists.
