Effect of Covid-19 Shutdown on worker well-being

published December 31, 2021

Goal

To gather insights on worker well-being from survey results within a mining company.

Preamble

The aim of the analysis was to identify any issues at the different mine sites and throughout the company brought about by the changed working conditions occasioned by the SARS CoV-2 pandemic.

Data

This survey consisted of both multiple choice and free-text responses, and was the latest annual survey conducted each year at this mining company. The survey was particularly focussed on the issues that arose (if any) while working through the SARS CoV-2 pandemic. The current survey instrument was designed as an online series of forms, accessible on the company intranet. The collected data was later extracted into an Excel workbook and prepared for analysis.

Method

There were a number of analyses run over the data.

1. Topic Modelling was run against the free-text field to highlight the main themes being discussed by the employees. inter-topic distance maps were generated to aid in determining the main topics discussed.

2. Sentiment analysis was conducted over the same free text fields to get an idea of the employee sentiment at each of the mining sites and within administration.

3. Sentiment scores were then correlated with each of the multiple choice responses to see if there were any stand-out issues.

Results

Three of the most prevalent topics brought up by the topic modelling exercise were:

1. Working from home (46%)

2. Collaboration (30%)

3. Site & health and safety (24%)

It was also found that during Covid-19, the administrative staff, who could all work from home, had the most positive experiences. Also, sentiment analysis together with the multiple choice responses highlighted more dissatisfaction from employees at two of the mining sites compared to the other sites.

So the whole exercise certainly assisted the company in pinpointing the issues of interest, and has helped them put in place fixes where needed.

What I found was really nice about this whole exercise was the requirement to merge exploratory analytics with machine learning to derive an idea about what messages the survey data was hiding. It was also with a sense of satisfaction the findings of the analysis were taken seriously and the company did indeed take appropriate actions to address employee concerns.