It was pretty cold today. I also noticed that my internet was pretty bad today. I remembered seeing online somewhere that temperatures could have an effect on cable networks. I decided to use a relatively small sample to test this hypothesis.
I recorded the high, low, and average temperatures for each day so far in January (18th) from WolframAlpha. I also recorded the downstream and upstream bandwidth from the first SpeedTest that day. Some days were skipped, but that was because I never felt the need to check the bandwidth because everything was running fine. I graphed this data with LibreOffice Calc. I noticed that the lines sort of matched up. I then used LibreOffice's CORREL function to determine a correlation between the downstream bandwidth and the average, low, and high temperatures. For the average temperature I got a correlation of 0.615, which shows a weak correlation, but a correlation none the less. With the low temperature I got a correlation of 0.577, which is a little weaker. With the high temperature I got a correlation of 0.383, which is much weaker than with the average. If anyone wants to see the spreadsheet let me know, and I'll upload it.
In conclusion, I believe that there may be some relation between temperature and the network speed.
I do realize correlation does not equal causation, and that these results are not all inclusive or even accurate to any degree. It's always fun to do a little bit of statistics, or is that just me?