World’s Top 25 Biodiversity Hotspots Losing Species Fast – What the Data Reveals
Description
This dataset showcases global patterns in biodiversity loss by highlighting the top 25 countries with the highest number of threatened species across three critical groups: plants, vertebrates, and invertebrates. Sourced from UNdata and last updated in November 2024, the data spans from 2004 to 2023. Each entry includes the country, year, and corresponding threat values, offering a comparative view of species vulnerability over time. The dataset has been cleaned, filtered, and categorized using Python (Pandas), with separate CSV files created for each species group. Visualizations are provided through interactive horizontal bar charts, enabling easy exploration via a year selector. This resource is ideal for conservation research, biodiversity education, and policy analysis focused on global ecological trends. Hypothesis ----------- We believe that countries with higher rates of industrialization, deforestation, and habitat disruption are likely to report a greater number of threatened species across plants, vertebrates, and invertebrates. By identifying the top 25 countries in each group over time, we aim to reveal global patterns of ecological stress that often go unnoticed in day-to-day policy decisions. Notable Findings ------------------- Some countries consistently rank high across all three categories, suggesting systemic biodiversity stress. Nations like Indonesia, Brazil, and India appear repeatedly in the top 25, highlighting the tension between economic development and environmental conservation. Interestingly, some smaller island nations show disproportionately high numbers of threatened species, underscoring their unique ecological vulnerability. This dataset also shows that invertebrates, often overlooked in conservation efforts, face significant threats in both developed and developing regions. How the Data Was Gathered ------------------------------- The data was sourced from UNdata, last accessed in November 2024, and includes annual records from 2004 to 2023. We focused on three key columns: country, year, and number of threatened species. Using Python (Pandas), we cleaned the dataset, filtered out continental aggregates, and separated countries into three species groups: plants, invertebrates, and vertebrates. For clarity, we selected only the top 25 countries per group for each year. The data was then structured into CSV files and visualized using Chart.js, making it easy to explore trends via an interactive year selector.
Files
Steps to reproduce
Methods and Processing: To recreate this dataset, begin by accessing biodiversity threat data from UNdata or other credible conservation databases such as IUCN or FAO. The focus should be on annual records of threatened species by country, covering 2004 to 2023. From the full dataset, extract key columns: Country, Year, and Threatened Species Count. Filter the data separately for Plants, Invertebrates, and Vertebrates. Remove aggregated continents and retain only individual countries. Finally, isolate the top 25 countries per year based on species count in each group. Software Used: Python (Pandas): For data cleaning, filtering, and CSV generation. Chart.js (JavaScript): For rendering interactive horizontal bar charts with color-coded visuals. HTML & Iframes: For embedding the finalized charts on the main I Hug Trees data pages. Jupyter Notebooks: For experimenting with python pandas and previewing data manipulation steps. How the Data Was Gathered, Cleaned, and Categorised: The data was originally collected from UNdata, with the last retrieval in November 2024. It included hundreds of country-year combinations across three major species types. After importing the data into a Pandas DataFrame, non-country entries (e.g., continents or regions) were removed. Next, species data was categorised into three distinct groups: plants, invertebrates, and vertebrates. For each group, the top 25 countries per year were sorted and saved into separate CSV files. These files were then hardcoded into JavaScript for fast rendering, avoiding real-time processing delays on the live site.