Elevate Analytics with ChatGPT
With the recent surge of generative AI, it is time to rethink the possibility of business analytics. As a practitioner in the field, I believe there are several themes where analysts can leverage tools like ChatGPT to elevate their work to the next level.
Ideation
Oftentimes, a data analyst’s work starts from a business question or hypothesis. Why has revenue declined? Will changing this product feature increase engagement? The list goes on and on. An analysts’ job is to brainstorm relevant metrics, or design an experiment to validate this idea. This process is highly influenced by personal experience and susceptible to bias. However, with ChatGPT, one can easily ‘outsource’ this work by asking it to brainstorm the answers first. For example, using prompts: [List 10 metrics/dimensions/frameworks to analyze revenue decrease.] [Design a controlled experiment on customer engagement, including recommendations on hypothesis, decision metric, sample size and test duration.] Analysts can receive preliminary answers to validate and refine, rather than starting from scratch.
Consultation
If we are being honest, a significant portion of time in analytics projects is spent on coding and data preparation. With project timelines often fixed, the time remaining to uncover meaningful insights is limited. ChatGPT can streamline this process by expediting the preparation further. Simply pose queries such as [Optimize this SQL query/table/database design with this code (code example).] [Troubleshoot this SQL/Python/Tableau error (error message) with this code (code example).] This minimizes the need for extensive searches on platforms like Google or Stack Overflow. As a result, it frees analysts from the mundane coding exercises, enabling them to focus more on outcomes than on specific methodologies.
Summarization
Effective documentation is a key quality for seasoned analysts, which includes codes, dashboards, models and data insights. However, the reality is a through documentation is time consuming. Leveraging ChatGPT we can bridge the gap between siloed information and structured documentations easily. [Write annotations on this code/model/dashboard. (Example 1, 2, 3)] [Write a summary or tell a data story with these insights. (Insight 1, 2, 3)] Now you can spend more time on the analysis rather than the paperwork. This also applies to preparing presentations, emails and other communications.
Education
Beyond day-to-day inquires, you can treat ChatGPT as a versatile resource when it comes to learning new domains. Thinking about the time before Gen AI tools becoming available, we would go to a website (usually starts with ‘Wiki’ or ends with ‘pedia’), and read through the definitions to learn a new term. They are comprehensive and through indeed, but lack of customization by industries or job roles. Now within your fingertips, you could have immediate access to a ‘private tutor’ to learn everything. [Explain this concept and include examples as if I am in this job role/industry.] [Recommend an agenda/book/course to learn about this subject with assessment questions.] Learning is a never-ending journey in the data analytics world. Gen AI tools like ChatGPT can flatten ‘the curve’ by creating content tailored to individual needs.
If you abstract data analytics as ‘the process of identifying patterns from numeric information’, then generative AI is ‘the tool to understand, reason and create content from various input based on user requirements.’ Comparing traditional analytics to generative AI is akin to ‘riding a bicycle’ versus ‘driving a racecar’. As these tools evolve rapidly, data analytics could soon become a relic of the past. In order to stay competitive as an analyst, one should think about the aspect of their job that can be replaced, versus where they could truly make a difference.
