Insights on Data and Analytics by Araf Khan

 


What are the leadership skills needed to drive effective change in organizations to achieve data and analytics excellence?


It’s unlikely that the whole middle or top management would be aligned with all the analyses & recommendations. In this case, a data analyst must step up and combine business intuition with sound analyses to push forth the optimum strategy. A data analyst represents not only the analytics team but the logical findings from any data set.


On the subordinate side, whenever an analytics team starts to roll, some routine tasks need to be automated or delegated. Even so, any critical thinking at any step of analysis can produce excellent insights. It is the data analyst's job to create such a culture and help the subordinates grow in that dimension. These two types of leadership skills are fundamental for any analyst.


Can artificial emotional intelligence be a substitute for real emotional intelligence?


From an emotional point of view, I would say no. Human emotion is complex consisting of layers within layers. We may be able to replicate the surface level of emotion with an artificial one but in that, the depth of complexity will remain absent. By surface level of emotion, I mean the facade of emotion without the absolute raw meaning behind it. 


Suppose, I am sad and the surface level of this emotion entails the reason behind it, the probable solutions and the associated expressions (all of these are logical and can be artificially built). But the raw meaning of that sadness is the depth of this emotion which incorporates the collection of my life experience and this particular emotion’s association with that experience, the collective outcome of human evolution and the role of that emotion (sadness) in such evolution, the seemingly contradictory reasons behind the sadness. All of these require the subjective notion which is not easy to replicate.


What does sales analytics mean?


This is just fascinating! Suppose your company makes & sells cookies to retailers. You are the sales manager. You take into account the sales trend, seasonality as usual. However, sales analytics happen when you take a scientific approach to the full sales journey. You analyze store location, size, assortment, footprint, neighbourhood etc. to make a complete profile of each store. Then you toss in each store manager’s (or owner’s) complete bio to make an index of each store based on multiple dimensions. Armed with this initial info, you then count your salespersons’ interaction with the store managers. Record the sales pitch. Analyze both the voice & clip. Find out what works & what doesn’t. Do it for a sample and then disseminate the best practice to the whole team. Continue to improve. Boom! You got full sales analytics.


What role does analytics play in measuring the engagement level of employees?


People analytics can save any company a lot of time and reduce much of the tensions & conflicts. Companies can tap into the plethora of their employee data and make that a holistic picture by incorporating every single feedback & performance review. The HR Analytics can then suggest a particular career path for each individual employee with timely tips & tutorials. This way, an employee doesn’t feel lost and becomes more engaged with the company.


How can data scientists and analysts improve their communication skills?


Very Important! Data Scientists can have loads of tools upon their sleeves but most of their outputs will fail if they can’t communicate. I’ve seen a lot of times long-term data-science projects being shut down only because the project champions couldn’t properly communicate the business impact.


As per the improvement, I personally think the main barriers in this communication gap are the overuse of jargon and unnecessary focus on execution. For example, you can use “classification/k-means/random forest” to indicate a classification problem. However, a simple term like “divide into different cohorts/groups” can easily communicate the same thing with less technicality. Also, it’s really important to communicate the summarized impact or use case of this classification rather than explaining the execution steps.


If any data scientist is aware of these 2 issues when she is communicating, much of the problem will be solved. Whenever any data scientist interacts with anyone from another team/domain, he can greatly improve his communication skills just by thinking that he’s talking to enthusiast standard 5 children. 


Insights by: Araf Khan

I’m currently working as an Assistant Manager - Analytics at Shohoz Limited, one of the fastest-growing startups in Bangladesh that embraced the superapp concept. I mainly deal with operational analytics focusing on our rideshare vertical. Before this, I worked as a Data Analyst for an e-commerce company where I had to implement a data-driven decision-making strategy for the whole value chain.