I have a SEO keyword tool (web application) that is currently under development. It scrapes and visualizes Google Trends data for large number of keywords based on some pre-defined criteria.
I would like to add a function to this tool that estimates the absolute number of searches per keyword, based on the chart data provided by Google Trends. The timeseries data is adjusted and normalized, so we need to find a way to de-normalize it: [url removed, login to view]
By zooming in close enough (looking at a small time interval like 1 hour), the chart usually gets coarse-grained and you may even see spikes that correspond to distinct search requests. This should enable us to create an algorithm that parses through different time ranges, combines them mathematically and returns an estimate of the number of searches for a given period.
Your task is to create an algorithm that uses as few queries as possible (there's already a working scraper providing the raw data). Ideally, you should also be able to implement the algorithm yourself or help another freelancer implementing it.
See [url removed, login to view]
[url removed, login to view]