Did I Choose the Right Business Analytics Program?
Wondering if a Master’s in Business Analytics degree can prepare you sufficiently for your future jobs? Let’s text-mining the job market
This project is a collaborative work of Jingnan Wang, Kuanglin Zhang, and me. Our team is the runner up of 2022 Fuqua Winter Data Competition.
Introduction
According to U.S. Bureau of Labor Statistics, the employment of management analysts, including business analysts, is expected to grow 14% from 2018 to 2028. As a result, graduate Business Analytics programs become one of the most popular graduate programs. Business Analytics bridges the gap between business and data analytics. Thus, it requires a wide spectrum of skillsets.
Graduate Business Schools’ top priority is to help students develop necessary skills for their future jobs, ensuring that students can find ideal job positions and advance in their careers in the long term. If there is a wide gap between the course content and the knowledge required by future employments, students may lack the motivation to study for that class. Also, classes that do not provide value for future jobs will take up time that could be spent doing job searching and gaining industry knowledge. In a relatively new industry like Business Analytics, companies’ requirements for a Business Analyst are constantly changing and it is challenging for Business Analytics Graduate Programs to design the curriculum so that the courses fit the demand of employers.
That is why we decided to investigate whether the Master of Quantitative Management: Business Analytics Program at Duke University (the program that we are currently studying in and abbreviated as MQM) teaches the skills that we will apply in our future jobs.
We hope to excavate the skills that are either overlooked or overtaught in MQM curriculum, and we will also analyze the reasons why certain skills required in jobs are not provided in MQM curriculum. Recommendations will also be provided to improve the curriculum design so that students can be better prepared for their future jobs. We hope that our analysis can provide our fellow MQM curriculum designers with precious insights to assist its future development.
Data Collection
Demand side
We analyzed the demand side of the job market based on job ads on Indeed.com. Indeed provides an official API for users to optimize their job research process using Python. After downloading the package indeed-python, we used the following code to acquire job titles and detailed descriptions.
We acquired 250 job ads located in the US with the keyword “Business analytics.”
Supply side
For the supply side, we focused on the MQM program. We collected course descriptions from Dukehub and syllabus from Canvas for every core or elective course. The final dataset includes 28 courses and detailed text descriptions.
Text Mining
The goal of our text mining model is to identify and quantify the business analytics skills demanded by the job market and taught in the program. We used WordStat 9.06 to build our model. WordStat provides a user-friendly Graphical User Interface and supports R/Python scripts and Tableau workbooks.
Our model was inspired by Seal et al. (2020). We used the following concepts to conduct text mining:
- document: each job or course description
- terms: words, n-grams (combination of adjacent words), and rules in documents
- dictionary: the dictionary documents how terms correspond to each category, which in our case, the skill.
After exploratory analysis, literature review, and web searching, we identified 12 soft skills and 21 analytical (technical) skills. The soft skills include “business acumen,” “communication skills,” “project skills,” “team skills,” etc. The analytical skills include “database programming”, “data visualization,” “machine learning process,” “programming language,” “statistics,” etc. In the dictionary, each skill consists of several terms with similar meanings. The dictionary-based text mining model can identify which skill is mentioned in each document and integrate results from different documents to calculate the percentage of documents that a specific skill category appears.
As we gather the percentage data of job descriptions and MQM courses to include a specific term respectively, we are able to analyze the results to compare the demand side (job skills requirement) and supply side (skill buildup in MQM curriculum). In addition to discussing the skill buildup in MQM courses and comparing it with skill requirement in the job market as a whole, we strive to further differentiate the MQM courses into core courses and elective courses. Also, with the performance of other 49 graduate analytics programs extracted from Seal’s research, we aim to compare MQM’s performance with it. In this analysis, we will utilize data visualizations to compare demand side performance with MQM curriculum’s performance on skill preparation and also compare MQM curriculum’s performance with performance of other programs.
Analyzing Results
Core courses
As we sorted the bars representing job cases based on its percentage in descending order on our dual-axis box plot, we discovered that the top five skills all fell into the soft-skill category rather than analytic skills suggesting a high demand for soft skills in real-life job cases. Specifically, Communication skills and Team skills as two prominent categories show appearance on over 60% of job descriptions (Communication Skills: 71.53%, Team Skills: 61.31%). The percentage seems reasonable that workplace collaboration between employees has been exceedingly common and demanded to complete a complex and time-consuming project and team skills and communication skills are extremely demanded in a collective effort. MQM Core courses seem to perfectly supply the need of soft skills and they show huge emphasis on communication as the core courses’ percentage to include Communication skills reaches 81.82%, even larger than that of the job cases. Every student in MQM program should fulfill the core courses regardless of their tracks, and core courses enable students to engage in in-class projects with students from other tracks to aggregate skillset from different areas. Most of the analytical skill categories appear in less than 25% of the job cases. It is possible that firms hope that their candidates learn and master analytical skills in a practical working environment with real projects, after they enter their jobs. For certain skill categories, we found huge gaps between the bars of job cases and MQM courses. For example, Decision Science, Big Data and Programming languages show higher percentages in MQM core courses than the job cases which potentially signifies that MQM core courses incorporate teaching in these analytical skills more than what the jobs require.
To better analyze the skills that are taught in MQM courses and their extent of matching the jobs’ requirements, we further constructed a scatter plot (Figure 2) to plot each skill as a point with the percentage of MQM core courses and percentage of job cases as vertical and horizontal axes respectively in a 2-D plate. As the plot shows, the shaded area is bounded by the lines representing the first and third quartile of the percentage of job cases and MQM core courses. Generally, we labeled the areas by “Matched”, “Missed” and “Over Taught” to explain the extent for a graduate analytics program to match the job requirements on a particular skill. If a program has most of its skills located in regions of “Over Taught” and “Missed”, it suggests that there occurs a mismatch between the skills taught in graduate programs and the job requirements. From Figure 2, we could see that skills such as Team Skills and Data Mining are within the area of matched skillset. At the same time, we can see that skill such as Database Programming is located in the “Missed” region, suggesting that MQM Core Courses lacks elements of a skill that is demanded by the emerging trend of industry. The same applies to Business Process and Cloud Infrastructure in the same “Missed” region. Spreadsheet is also located in the “Missed” region. The lack of emphasis on the Spreadsheets in MQM Core Courses is probably because graduate analytics programs focus more on the teaching of coding languages that most students do not own prior experience on. Although spreadsheet seems to be a basic skill that most people possess, they sometimes overlook its importance in daily work environment. In comparison, we found that certain analytical skills such as Forecasting and Programming Languages locate within the “Over Taught” region and possibly MQM curriculum designers should reconsider whether these analytical skills are overly stressed or lack commonality and applicability in this dynamic business analytics industry. In short, the core courses seem to have a number of points outside of the “Matched” region and have the room for further improvements.
Elective courses
Moving on to analyze MQM electives, we constructed bar plot and scatter plot Figure 3 and Figure 4 respectively to help us analyze the correspondence of skills. From Figure 3, we observed that both MQM Electives and job cases regard Communication Skills and Team Skills as top priority skills with percentage importance. In comparison, skills such as Database Programming and Information Technology are placed with more importance in job cases than in MQM electives. One skill that is noticeable is Business Process. In MQM core courses, Business Process presents as a skill that lacks teaching to a great extent since its percentage of job cases (52.92%) largely exceeds the percentage of MQM Core Courses (18.18%). In electives, however, we saw better performance on MQM in meeting the demand of job cases as both the percentages of MQM electives and job cases are around 50–55%. The reason behind the difference of performance in electives and core courses is largely due to the difference in course design. MQM is a Business Analytics program that aims to teach students practical analytics skills in a business setting. Therefore, MQM requires every student to handle necessary analytics skills such as programming skills and places most of its analytics courses as core courses since these technical skills will apply to any business environment. Considering that different students may have different career interests and professional directions, MQM offers choices of tracks (Forensics, Finance, Marketing, Strategy) in order to fit students with their dream career route. As a result, these career-oriented electives are designed to incorporate diverse business settings and the diversity explains why MQM electives supplement the insufficiency of MQM core courses to meet the job demand on the skill of Business Process.
In Figure 4, we noticed that most of the points are located in the “Matched” areas, suggesting that the diversity of skillsets taught in MQM electives largely match what future jobs need. It is still worth mentioning that skills such as Database Programming remain to in the “Missed” region and it can be a sign that both MQM Core and Electives do not cover the necessary amount of training of this skill and need revision of curriculum.
Merging core courses & electives together
Lastly, we merged all MQM electives and core courses together to analyze the performance of how MQM courses match the skillsets requirements of real-world jobs. From Figure 5, we can see that MQM perfectly develops the necessary soft skills such as Communication Skills and Team Skills stressed by modern world business and workplace culture. Meanwhile, we noticed that MQM curriculum lacks teaching in Database Programming and Information Technology since we observed little presence of these skills in MQM courses. However, we also observed some biases that do not accurately reflect the reality at MQM. For example, the plot shows that visualization tools exhibit much higher percentage in job cases (24.09%) than in MQM course (10.71%). However, as MQM students, we have opportunities to conduct data visualization using Tableau & R in nearly every course. Possibly, the percentage of MQM data visualization courses is underestimated because Visualization Tools are specified in the course syllabuses rather than the names of the courses. The same applies to Business Acumen as lots of business terms deliver the same meaning as Business Acumen and possibly reduce the instances of finding Business Acumen as a unique term in course syllabuses.
We further looked at Figure 6 to excavate the insights behind the performance. We indeed noticed that most of the points locate within the “Matched Region”. We focused on the several points that are not located within the “Matched” region. For example, points of certain technical skills such as Data-Driven culture, and Deployment Techniques are located in the “Missed” region suggesting that MQM Courses insufficiently supply the skillsets demanded by job cases. Indeed, MQM is a program that lasts only 10 months, and that it does not require perfect mastery of technical skills for admission. Due to these reasons, the curriculum designers reasonably find it short to cover all the technical skills that different jobs require for MQM students. They can only make the curriculum cover as many topics as the program length allows. We also found that skills such as Project Skills are located in the “Missed” region suggesting that MQM should help students polish these skills more to better satisfy the need of future jobs. Perhaps the extra time needed for polishing could come from the reduced time of polishing skills that MQM overly stresses such as Forecasting and Decision Science which are within the “Over Taught” region”.
Comparison between MQM and other graduate Business Analytics Programs
How do MQM programs compare with other graduate Business Analytics Programs in terms of meeting the supply of skills of real-life jobs? For this question, we plotted Figure 7 which suggests the performance of other graduate analytics programs in meeting the skillset that jobs require, and we aimed to compare Figure 6 and Figure 7 to excavate insights from the distribution of points on the two plots. Compared to Figure 7 which has its points distributed diversely with no pattern on the plot, Figure 6 has most points in the “Matched” region suggesting a better performance of preparing students for the skillset that they really need for their future jobs. This is to be expected since MQM is one of the best graduate Business Analytics programs in the US. However, better performance does not equal perfect performance. MQM curriculum setup can be improved in certain directions based on our research results and we are ready to provide recommendations.
Insights & Recommendations
Based on the analyses above, we obtained the following insights into the skills taught by MQM:
- In terms of soft skills, MQM perfectly develops students’ communication skills but does not develop students’ project skills sufficiently.
- In terms of technical skills, MQM does not prepare students very well in the field of Cloud Infrastructure and Database Programming. On the other hand, MQM emphasizes Decision Science, Big Data and Programming languages more than what the industry requires.
- The diversity of skillsets taught in MQM elective courses incorporates different business settings and meets the job demand on the skill of Business Process, which supplements MQM core courses perfectly.
We think the following could help MQM better prepare students for their future jobs:
- Reconsider the amount of course time allotted to Decision Science and Big Data since these courses are not as emphasized in real-world jobs. Also, these courses require a substantial amount of time to study in-depth. Rather than having a basic understanding of these advanced concepts, it would be much more useful to assign more time for courses like Cloud Infrastructure and Database Programming, which are very much in demand in the industry.
- Remain acute towards the development of Business Analytics industry. Set up a mechanism where course concepts that are gradually becoming irrelevant in the industry will be removed and course concepts that nurture skills that are becoming increasingly in demand in the industry will be added on an annual basis.
- Allow students to take elective courses not only in the Business School, but also in other schools at Duke University to enrich their knowledge in the industries that they are interested in. This diversity of course selection can further supplement the core courses at MQM (as we have already seen from the elective courses at MQM) and help students polish their skills.
As one of the fastest-growing industries, the Business Analytics industry is also constantly evolving, which makes it challenging for graduate-level Business Analytics education to implement the most up-to-date courses that perfectly match what the industry is looking for. Though delivering on most of the skills required, there are still classes that can be implemented to further enrich students’ skillsets for the industry. We hope this research can provide guidance on how graduate Business Analytics programs can improve their curriculum and look forward to the bright future that lies ahead for the Business Analytics industry.
Resources
Interactive visualization: Did I Choose the Right Program? | Tableau Public
Code and data: yijiemkii/-Right-Program (github.com)
Reference
Kala C. Seal, Linda A. Leon, Zbigniew H. Przasnyski, and Greg Lontok, Delivering Business Analytics Competencies and Skills: A Supply Side Assessment (2020), INFORMS Journal on Applied Analytics 2020 50:4, 239–254
