Hong Kong River Quality



This project aims to answer the questions: “Are rivers in Hong Kong clean?” and “Which river is the cleanest”.

Hong Kong as a small island city is actually surprisingly large. The territory boasts roughly around 200 small mountain streams and a couple dozen or so major rivers. These major rivers are routinely monitored for their water quality by the Environmental Protection Department (EPD) of Hong Kong since 1986.

Data Source

Data: https://data.gov.hk/en-data/dataset/hk-epd-riverteam-river-water-quality-historical-data-en

Key: https://cd.epic.epd.gov.hk/riverpsi/historical_river_data_dictionary_en.pdf

The EPD collects a very comprehensive range of water parameters. Many of which are only meaningful in some specific scenarios. In the next section I’ll be going through a few of them that interests us and explain why.


For the purpose of this analysis I’ll be looking into the following parameters below. Along with a brief explanation why.

  1. 5-Day Biochemical Oxygen Demand (mg/L), or BOD5 in short, measures the amount of oxygen used after a 5-day incubation period. What this represents is the amount of organic matter in the water. The lower the better.

  2. Chemical Oxygen Demand (mg/L), or COD, is very similar to BOD5 but the incubation process is sped up using some chemical processes. This broadly represents the amount of foreign chemical substance in water.

  3. Conductivity (μS/cm) -

  4. Dissolved Oxygen (mg/L) - DO doesn’t really represent how clean water is but the higher the DO, the better the water’s ability to purify itself by sustaining good biodiversity.

  5. E. coli (counts/100mL) - This and Faecal Coliforms basically indicates if the water has poop 💩 in it. E. coli is one of the various coliforms present in faeces.

  6. Faecal Coliforms (counts/100mL) - This measure the concentration of 💩 bacteria in the water. A high level indicates that untreated sewage/adventurous outdoor toilet goer may be leeching into the river at some point.

  7. Oil and Grease (mg/L) - Also known as FGO (fats, oils, grease). This is one of the main indicators in waste water treatment that helps quantify whether anthropogenic water is treated properly. Waste water from kitchens (and people in general) are high in FGOs.

  8. Total Solids (mg/L) - This measures the amount of solid particles in the water

  9. Suspended Solids (mg/L) - This measures the amount of tiny particles that are too small to sink in the water column.

  10. Turbidity (NTU) - Turbidity is a fancy way of saying clarity. The cloudier the water, the more turbid it is.

Data analysis

View the code natively on Kaggle… or read on!

Data cleaning

Data wrangling


Chi Kit Yeung
Chi Kit Yeung
Jack of all trades, Master of some

Aspiring Data Scientist obsessed with nature.