Data Representation

1 day = 86400 seconds.

If we are sampling the environment at 35 frames per second (vision-based). Then we would have 3024000 records in a day. This would mean that we have 90720000 records in a month. These would be just the rows of a table. The columns could be enumerated upto a large scale as well. Thus, the dimensions of a matrix/table would be really big.

To represent data and encode it in a hard storage file, it is essential to build a hierarchy that allows optimal computation. Advanced data analysis require data spanning multiple years of records. Even if these files are saved on a monthly basis, it would still yield in files which cannot be analysed in real-time.

In terms of data warehousing, the larger the warehouse, the more time it would require to save and access objects within it. Especially in a structured format.

Conclusion: Since the data is based on a temporal values, we can associate it with time periods. This allows us to divide the data into months and days eventually. Databases can be built according to daily acquisition of data.