Istanbul Water Crisis Prediction
For this prediction, I will use the data provided by the Istanbul Municipality. These are basically:
-Dam Occupancy Rates Of The Last 6 Years For 12.21.202X
-Total Annual Rain
-Annual Amount of Water Pumped From Surrounding Provinces
-Annual Consumption
I divided all water data into “10 Million Cubic Meters” to make the data simpler. In this way, our graphics will be more understandable.
However, it is very difficult to interpret these data directly with each other. For example, in a year with little rain, too much water may have been pumped from surrounding cities. Or, in a year with excessive water consumption, a lot of rain may have come.
More importantly, none of these directly determine the occupancy of the dam. Unlike internet packages, the amount we do not use at the end of the month is not reset at dams.
Therefore, we will basically consider the 1-year occupancy changes of dams, not dam occupancy rates. In this way, we can directly link the change in dam occupancy rates in a year to annual consumption and annual total water input.
Here, it is useful to remind that, in the last 6 years, no water source has been added to Istanbul. So each percentile represents the same amount of water.
Now is the time to plot the “Net Water Balance” and “Percentage Of One Year Change In Dam Occupancy.”
-The first is an outlier in 2016 (Net+9): A mistake in the official figures may be due to malfunctions, leaks or factors that we cannot calculate. It may also be of interest to the Political Crisis in Turkey in 2016. But it’s an extreme result compared to the other three. So I’m going to take out 2016 while I’m designing artificial intelligence.
-The second important problem is the slowdown in dam occupancy rates when the net water inflow to the city exceeds 20. This again can be related to various problems. Maybe it’s an annual thing. But the most logical theory that comes to mind may be that more water ingress than the dams of the city can store makes no sense. Yes, water is coming, but it cannot be stored. After all, the number of rainy days in the city is clear. Floods may not contribute to dams. There seems to be a saturation point above “Net +20”.
I think, i made the necessary arrangements. It is time to turn our data into a prediction algorithm.