Weighted Average Download Speed at the Country level

Contents

Weighted Average Download Speed at the Country level#

Using Ookla’s opensource dataset we calculate an average download speed weighted by population for each country quarterly from 2019 till 2nd quarter of 2025.

Methodology#

  1. Filter Ookla’s data by country.

  2. Aggregate Ookla’s data to the desired zoom level. The zoom level should be smaller than 16 which is the level at which Ookla delivers their data. This step is optional if you want to work with 16 agregation level

  3. Calculate an average speed for each tile by averaging the speeds at lower level tiles. Tiles with no speed as disregarded from the average

  4. Aggregate Worldpop data to each tile, this results in an estimate of population at each tile.

  5. Calculate an average download weighted by population as follows

\[ WeightedAvgSpeed_{tile_{i}} = \frac{Pop_{i} * AvgSpeed_{tile_{i}}}{\sum_{i=1}^{m}Pop_{i}} \]

Note: The data processing for this was done on the ookla-data-processing notebook.

Results#

In the following plot we can observe 3 groups of countries by their development of internet infrastructure:

  • Hige level: ARE, KWT, QAT

  • Mid level: BHR, JOR, OMN, PSE, SAU

  • Low level: AFG, DJI, DZA, EGY, IRN, IRQ, LBN, LBY, MAR, PAK, SYR, TUN, YEM

In order to follow the speed of development at each country, we also plotted the change in the Weighted Average Download Speed across the time. The change was calculated as follows

\[ Change = \frac{WADS_{t} - WADS_{t-1}}{WADS_{t-1}} \]
Contents