Page 175 - Contributed Paper Session (CPS) - Volume 3

CPS1983 Chong N. et al.
Through the use of ACS, time spent on manual classification is reduced and
resources can be further deployed more effectively.

3. Results
The implementation of these initiatives has helped improve operational
efficiency and data quality. Before predictive modelling was implemented,
interviewers were unaware of the optimal timings to contact respondents and
as a result, the successful completion rate for fieldwork was only around 20%.
This proportion has more than doubled after the implementation of the model
as it provides interviewers with the knowledge on when and how to contact
the survey respondents.
The completion rate is further boosted to around 60% with sentiment
analysis as case assignment is done more strategically, by allowing more
experienced interviewers to handle difficult respondents. Sentiment analysis
has also helped interviewers to improve on their interviewing skills.
After the data is collected, ACS has helped to reduce the time needed for
data verification. More cases can be verified to ensure data consistency and
robustness with the same amount of time. More importantly, ACS eliminates
the human variable factor when classifying the occupational information into
their respective codes. This ensures consistency for each code and eliminates
any subjectivity on the part of interviewers.

4. Conclusion
With increasingly widespread use, data analytics will emerge as an
important tool in the statistical production process. The three initiatives
described in this paper is the start of revolutionizing the way official statistics
is produced in future. Encouraged by the positive results, MRSD will continue
to explore new initiatives in data analytics to aid in its survey operations.

References
1. James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An Introduction to
Statistical Learning: with Applications in R. New York: Springer.
2. Manning, C.D., Raghavan, P., Schütze, H. (2008). Introduction to
Information Retrieval. Cambridge (England): Cambridge University Press.

164 | I S I W S C 2 0 1 9