Insights into Financial Strength, IPO Myths, and Global Layoff Trends Using Data Analytics

Sigma
11 min readJan 14, 2024

--

By Sai Pranav, and Murali Krishna

Introduction

In the contemporary corporate landscape, layoffs have become significant, influenced by various economic, industrial, and company-specific factors. Our project explored a comprehensive lay- offs dataset, aiming to uncover patterns and insights through rigorous Exploratory Data Analysis (EDA). Leveraging a diverse set of analytical tools and libraries, such as Pandas for data manipulation, Seaborn and Matplotlib for data visualization, and Plotly for interactive plots, our analysis delved into the multifaceted aspects of layoffs. We employed statistical methods, including linear regression from SciPy, to interpret the data meaningfully. The utilization of advanced techniques like K-Means clustering and standard scaling further enriched our understanding of the dataset’s intricacies.

This report aims to present our findings, offering a detailed view of the layoff landscape across various industries and geographies. The insights drawn from this analysis are intended to pro- vide valuable inputs for stakeholders, including business leaders, policymakers, and economic analysts, offering a data-driven perspective on the patterns and trends in layoffs worldwide. This introduction sets the stage for your detailed EDA findings, highlighting the tools used and the scope of your analysis.

Problem Statement

The primary aim is to analyze and understand the patterns and trends in layoffs across various industries and locations. The focus is on identifying key factors contributing to layoffs and how different sectors and regions are affected. The analysis aims to provide a comprehensive under- standing of the layoff landscape, identifying industries and areas most affected, trends over time, and the potential impact of financial factors like funds raised. In this Project, We have a dataset consisting of 2103 entries for total layoffs, 2044 for percentage laid off, and 2820 for funds raised, indicating the presence of missing values in these fields.

Dataset and Attributes

Dataset link: https://www.kaggle.com/datasets/swaptr/layoffs-2022

The dataset is based on the EDA file, is centered around layoffs from various companies. It contains a range of attributes that offer insights into these layoffs’ scale, industry distribution, and geographical spread. Here’s an extensive explanation of the dataset and its attributes:

  • Company: This column lists the companies that have reported layoffs. It directly references the organizations affected and is crucial for identifying industry trends and company-specific issues.
  • Location: This attribute specifies the geographical location of the companies, typically given as city names. The location data is essential for understanding the geographic distribution of layoffs and identifying if certain regions are more affected than others.
  • Industry: The industry column categorizes each company into a specific sector, such as Crypto, Real Estate, Data, etc. This classification is vital for analyzing which industries are experiencing higher rates of layoffs and understanding sector-specific trends.
  • Total Laid Off: This numeric field represents the total number of employees laid off by the respective companies. This is a critical metric in the dataset as it quantifies the impact of layoffs at a company level.
  • Percentage Laid Off: Alongside the total number, this column provides the percentage of the laid-off workforce. This attribute offers a relative sense of the layoff’s impact, which is especially useful for comparing companies of different sizes.
  • Date: The date of the layoffs or the report gives a temporal context to the data. It allows for the analysis of layoffs over time and can help identify any temporal patterns or trends.
  • Stage: This column likely refers to the stage of the company, such as startup, growth, maturity, etc., though the specific categories need to be detailed in the provided excerpts. The stage of a company can be an essential factor in understanding its stability and the reasons behind the layoffs.
  • Country: This attribute specifies the country in which the company operates. Like the location, it’s crucial for analyzing the geographical spread and impact of layoffs on a global scale.
  • Funds Raised: This numeric field indicates the amount of funds raised by the company, presumably over its lifetime. This financial metric can be critical in understanding the finan- cial health of the company and its potential correlation with layoffs.

Data Cleaning and Preprocessing

The dataset you provided underwent several data cleaning and preprocessing steps to pre- pare it for analysis. Here is a detailed explanation of these methods:

Overview of the Dataset

1. Handling Missing Values:

1.1. Initial Assessment

  • To check for missing values in the dataset, we used the df.isna().sum() method.
  • The method revealed missing values in several columns, including:
    - ’total_laid_off’
    - ’percentage_laid_off’
    - ’funds_raised’

1.2 Dropping Rows with Missing Values

  • Rows with missing values in crucial columns were dropped from the dataset.
    - Columns: ’total_laid_off’, ’location’, ’industry’, ’stage’
df.dropna(subset=['total_laid_off', 'location', 'industry', 'stage'], inplace=True)

1.3 Filling Missing Values

  • For columns ’percentage_laid_off’ and ’funds_raised’, missing values were filled with the median of their respective columns.
    - Columns: ’percentage_laid_off’, ’funds_raised’
df['column'].fillna(median_value, inplace=True)

1.4 Data Transformation

  • Conversion to Datetime: The ’Date’ column was converted to a datetime for- mat using pd.to_datetime(df[’Date’]). This is important for any time series analysis, as it ensures that dates are correctly formatted and recognized.
  • Indexing by Date: The dataset was then indexed by the ’Date’ column to facilitate time-based analysis. This step is crucial for chronological analysis and time series forecasting.

Exploratory Data Analysis (EDA)

Density Curves:

The graphs display density curves for total layoffs, the percentage of employees laid off, and funds raised. The first graph shows that most companies have a low number of layoffs, with a sharp peak at the lower end. The second graph indicates a similar skew towards lower percentages of layoffs. The third graph suggests most companies have raised a relatively small amount of funds, again with a peak at the lower end, indicating that higher fund-raising is less common.

Linear regression

Based on the scatter plot and the accompanying statistical measures, there is a weak neg- ative relationship between the percentage of the workforce laid off and the funds raised by companies. The p-value is 2.24e-02 (or 0.0224), which suggests that the relationship ob- served is statistically significant at the 5% significance level (since the p-value is less than 0.05). However, the correlation is very weak, While the relationship is statistically significant, the correlation coefficient is so close to zero that it suggests no practical linear relationship. This indicates that within this dataset, the amount of funds a company raises is not a strong predictor of the percentage of its workforce that may be laid off

Count of companies by industry

The graph depicts the number of companies by industry, with the retail and transportation industries having the highest counts. In contrast, AI and Manufacturing have the lowest, indicating the industry-wise distribution of companies in the dataset.

Count of companies by countries

The bar chart shows the number of companies by country and displays the top 15 countries. The United States leads substantially, followed by India and Canada. This illustrates the dominance of the U.S. in this dataset and highlights the global distribution of companies, with representation across North America, Europe, Asia, and South America, indicating the diverse multinational presence of companies considered in the dataset.

Layoff Trends in top 5 sectors

The line graph illustrates layoff trends over time in three-month intervals for the top five sectors. It shows significant layoff fluctuations, with peaks and troughs indicating varying levels of layoffs in different quarters. Transportation and Retail sectors exhibit the most pronounced changes, suggesting seasonal or cyclical impacts, as well as potential market- driven events affecting these industries’ workforce numbers

Monthly Layoffs

The graph represents monthly total layoffs with a significant spike, indicating a peak in lay- offs at a particular point in time. The blue dashed line represents the best fit line, showing the overall trend excluding the outlier. Despite the peak, the trend line suggests a relatively stable pattern of layoffs over the observed period, indicating that the spike may be an anomaly rather than a consistent trend

Total layoffs by industry

The bar graph displays total layoffs by industry, revealing a descending order of impacted sectors. Retail leads with the highest layoffs, followed closely by Consumer and Other categories. The trend shows a steep decline in layoffs as industries transition from Transport to AI, suggesting that some sectors are more vulnerable to layoffs than others, possibly due to economic, technological, or sector-specific challenges.

Pivot Table

It aggregates data by summing up layoffs, with missing values filled with zero. The table is then styled to format the numbers without decimal places. This pivot table allows for a clear, concise comparison of layoffs across different industries and countries, highlighting areas with higher and lower job losses.

Top 5 sectors with top 5 countries

The bar chart illustrates layoffs within the top five sectors across the top five countries with the highest layoffs. The United States has the most significant number of layoffs across almost all sectors, particularly in the Consumer sector. The chart demonstrates the variation and distribution of layoffs by sector within each country, with other countries showing notably fewer layoffs compared to the U.S. across corresponding sectors.

Correlation Matrix Heatmap

The heatmap represents a correlation matrix, visualizing the relationship between total lay- offs, percentage laid off, and funds raised. The color intensity indicates the strength and direction of the correlation. Dark blue shows a positive correlation, while dark red indicates a negative correlation. The matrix shows very weak correlations between the variables, suggesting that within this dataset, these variables do not have strong linear relationships with each other.

Total layoffs across USA

The bar chart displays total layoffs in various sectors within the United States, with the Consumer sector experiencing the highest number of layoffs, followed by Retail and Other. This visual representation emphasizes the disproportionate impact on the Consumer sector compared to others, like AI and Aerospace, which had the least number of layoffs. This data could suggest sector-specific economic challenges or transformations impacting employment.

Mean percentage of layoffs:

The bar chart visualizes the mean percentage of employees laid off across various indus- tries. Healthcare and Food industries have the highest mean percentage of layoffs, sug- gesting a significant workforce reduction in these sectors. The bars represent the average layoffs per industry, providing a quick comparison of how different sectors have been af- fected. Shorter bars for industries like Legal and AI indicate a lower mean percentage of layoffs in these fields.

Pie chart

The pie chart illustrates the distribution of total layoffs by company stages, with a dominant share of layoffs occurring in post-IPO companies, accounting for 53%.

Clustering

The bubble chart categorizes companies into clusters based on the total number of layoffs, with the size of each bubble reflecting the layoff magnitude. Blue represents companies with low layoffs, red for medium, and green indicates companies with significant layoffs. The chart highlights that while most companies have a low to medium number of layoffs, a few companies, particularly those represented by larger green bubbles, have experienced a disproportionately high number of layoffs.

Worldmap

The world map visualizes total layoffs by country, with color intensity corresponding to the number of layoffs: darker colors indicate higher layoffs. The United States stands out with the highest layoffs, followed by various shades across other countries, reflecting lower num- bers. This map highlights the global impact of layoffs, showing that while some countries experience high unemployment rates, others are less affected, providing a geographical perspective on the economic challenges faced worldwide.

Top 20 companies by percentage laid-off

The bubble chart represents the top 20 companies by the percentage of employees laid off, with the bubble size indicating the layoff rate. Larger bubbles denote a higher percentage of workforce reductions. The chart is color-coded to differentiate between companies and showcases a range from the highest percentage at the top to lower percentages as the bubbles decrease in size. This visualization highlights the companies most affected by layoffs in a comparative context.

Top 20 companies by total employees laid off

The bubble chart displays the top 20 companies by total number of employees laid off. The size of each bubble corresponds to the total layoffs, with Amazon and Meta having the largest bubbles, indicating the highest number of layoffs. The chart offers a visual com- parison among leading companies, showing a significant variation in the scale of layoffs, which decreases markedly among companies with smaller bubbles, like IBM and Groupon, towards the end of the chart.

Mean percentage of total laid off

The bar chart illustrates the mean percentage of employees laid off in the top 15 countries. South Korea shows the highest mean layoff rate, significantly more than the other countries

listed. The chart displays a descending order of mean layoff percentages, with countries like Israel, France, and Germany having the lowest rates among the top 15.

Total laid off- Top 15 countries

The bar chart shows the total number of employees laid off in the top 15 countries, with the United States far exceeding other countries in layoffs. India, the Netherlands, and Germany follow, but with significantly lower numbers. The bars are color-coded with the respective national flags, emphasizing the stark contrast between the United States and other coun- tries. The data suggests a considerable geographical disparity in the impact of layoffs on the workforce.

Conclusion

  1. The exploratory data analysis (EDA) on layoffs revealed significant insights. It identified a substantial difference between the means of funds raised and the percentage of employees laid off, suggesting a complex relationship between a company’s financial influx and its layoff decisions.
  2. The recent increase in layoff trends is because of multiple factors involving the COVID-19 pandemic, Russia, the Ukraine war, and the collapse of a few financial institutions. The instability of supply and demand in the markets, affecting the company’s revenue, led to budget cuts to maximize profits through layoffs. Through this extensive analysis, we will be well-informed in making career or higher education choices wisely. The business point of view is where we can have an idea of our wise investments in the sectors or the company.
  3. The analysis also highlighted industry-specific trends, where certain sectors experienced higher layoffs.

This study was crucial in understanding the dynamics of layoffs in the corporate world, providing a data-driven perspective on how economic and industry-specific factors influence workforce reductions. This understanding aids in better preparing for and mitigating the impacts of such challenging corporations.

--

--

Sigma
Sigma

Written by Sigma

The Business Club Of NITT

No responses yet