Inspired by the Density calculations in Theoretical Chemistry and Statistical Mechanics, molFrameR will make density statistics much easier to calculate visualize!
You may be familiar with R and ggplot2's density curves to represent frequency statistics, such as this one:
But, what if you want to describe the the plot by metrics other than frequencies? such as mean values?
| x | 1 | 2 | 1 | 1 | 3 | 3 | 1 | 2 | 1 |
| y | a | b | c | d | e | f | g | h | i |
Take the following table above as an example, let's say you want to bin all values such that the x values in the plot is 1, 2, and 3. But, instead of counting the frequency of the values, you want to take the mean of all values that correspond to x = 1 to give a mean value of all entries, so you end up with the table below:
| bins | mean(y) |
|---|---|
| 1 | (sum(a + c + d + g + i)/count(x = 1)) |
| 2 | (sum(b + h)/count(x = 2)) |
| 3 | (sum(e + f)/count(x = 3)) |
This is the problem molFrameR aims to solve by providing functions that will allow one to organize data in such a way without requiring SQL libraries to manipulate the data frames using keys and joins. molFrameR also contains other parameters that makes the resulting data more customizable.
Now that we know what molFrameR is used for, let's take a look at the bells and whistles!
Custom histogram bin sizing!
Yes, the bin function also allows one to declare how large the histogram sizes are to manipulate the resolution of the density plot.
SQL-like capabilities without requiring knowledge in SQL!
In addition, molFrameR contains functions that also allow one to compare multiple tables and join them based on conditions, i.e. creating a new table with entries from either data frames that are compared line per line. Quite similar to how SQL makes use of joins and unions.
Compatibility with packages such as DplyR and magrittR pipelines!
molFrameR functions can be included as part of data munging done with DplyR.
- change method names for generalization
- add overloaded methods with default values
- upload to R CRAN repositories
- implement other methods such as Monte Carlo sampling
- create a Python library counterpart