Why We Need More Great Data Analysts
I know, the title sounds a little bit contradictive, right? We are living in the golden age of Machine Learning. The computing power has developed a thousandfold. Nowadays, our smartphone is more powerful compared with the computer that helped NASA to launch their astronauts to the moon. There are hundreds of Machine Learning applications in the real world starting from self-driving cars, diagnostics applications for X-Ray results, protein structures prediction, etc. Yes, I agree we should invest more in the development of knowledge for breakthrough and forward-thinking in the real-world applications of Machine Learning to unlock the hidden value of the inventions for business and society.
However, the whole invention and findings would require a solid start. Those investments need to be very well justified. At the start of everything, a Data Analyst would have the most valuable role to provide context for the problems. A great Data Analyst is the one who would help you decipher the problems that are worth solving. They will fly beyond the data, find the needle in a haystack, and provide excellent recommendations to the team. Great Data Analyst is the leader in the battalion to guide the directions before Data Scientist, Data Engineer and Machine Learning Engineer come to the battle. Without them, the company will only have a group of experts doing impractical advanced analytics projects or even losing the battle before it all started.
The diagram above shows the interaction process of Data Analysts with other project team members such as Data Engineer and Data Scientist. Oftentimes, Data Analysts also provide data quality assessments to Data Engineers since Data Analysts really know about the data because they use it in daily life. This is a very important process where the quality of the model is strongly defined by the data themselves. The balance of the ecosystem needs to be achieved, it is incorrect to focus solely only on Data Scientist or Machine Learning Engineer. A healthy data ecosystem should comprise a great support system too.
Now, the next question is what differentiates a great Data Analyst?
1. Great storyteller
They have the wittiness to explain the problem in a wider approach. What I meant with the broader approach; they could explain the problem, so it became easy enough to be understood by the wider audience. Statisticians might dwell in the statistical difference and proving null hypothesis to provide valid conclusions. However, a great Data Analyst could do it in a manner that is intuitively easy to understand.
2. Speed
They are excellent with speed. These analysts know exactly what to look for in the data to explain the current phenomena that need to be understood. Data Scientist might dig thousands of features, but Data Analyst could do with “just right” number of features. Great Data Analyst uses data to minimize the chance that decision-makers will come to an incorrect conclusions. The Data Analyst understands what to answer quickly because they are well-tuned with their business stakeholders.
3. Reward vs Effort Ratio
They know how to balance the reward vs effort. A great Data Analyst would balance the reward and the effort without compromising the results. Sometimes regression analysis would do just fine to solve the problem compared with a neural network. Don’t overkill, we are not trying to kill mosquitoes with AK-47.
4. Context
Able to channel the internal data and external data and finding good conclusions. A great Data Analyst is a well-rounded person that has a broader view. They must be open-minded, flexible, and always questioning themselves. I have one experience that I could not forget, I did an incomplete observation that I just realized 6 months from the moment that I built the analysis. I made a conclusion which I said the more subscribers used communication apps their spending would increase. I saw the opposite when they don’t use the communication app their spending is down. I was half-blind. I should have known that when they didn’t use the communication app, probably they moved to competitors, and they became more active in the competitor’s networks. This is a huge lesson for me to always question myself.
Data Analyst is not a second-class citizen. A great Data Analyst is as valuable as Data Scientist, Data Engineer, or Machine Learning Engineer. Data Analyst uniquely fills in the gaps in Data Scientist, Data Engineer, and Machine Learning Engineer. I think the role of a great Data Analyst in the golden age of Machine Learning is stronger than ever. Thus, I have made my case why do we need more great Data Analysts. Please let me know your view in the comment sections. Thank you.
This post is dedicated to all Data Analysts in Telkomsel and Indonesia
Picture source: https://www.businessintelligenceinfo.com/data-mining/tibco-spotfire/the-5-stages-of-machine-learning-and-the-unique-data-requirements-of-each with an extra explanation from the author of this article
