1. Used Chat gpt but there was negative values.
2. clipped negative values in data to zero
3. replaced zeroes with random ints (1-3), those timings are from 1am to 5am, which really have random cyclists. (score= 0.67)
4. replaced zeroes with random ints (1-168), (score= 1.01)
5. replaced zeroes with random ints (1-6), used Long Short-Term Memory(LSTM). (score= 0.662)
6. replaced zeroes with random ints (1-20), used Long Short-Term Memory(LSTM). (score= 0.649)
7. replaced zeroes with random ints (1-40), used Long Short-Term Memory(LSTM). (score= 0.657)