Hello Sunaulla; Your project idea is nice and tidy. As a first attempt I'd do the following.
1. Find a place where the flower density varies. You can set up plots near one another.
2. During timed intervals you can count bee visits in the two plots.
That will give you an idea about the relation you are studying. It would also give you an idea about the potential problems with your method as well as the needed statistical analysis. With that in mind you can design a good study.
Yes , flower density and bee density could have positive correlation but more important is to see the level of pollination , fruit set and eventually the yield as a barometer of crop response via- a- vis bee density..
For measuring correlation between bee and flower density one needs few different measurements, usually 10 or more.
When it is measured during time (like during consecutive days) the correlation coefficient "r" may be used as a measure for the relative attractiveness of the specific flowers for the bees.
The attached graph shows seasonal density of flowers, honeybees and fruits on 'Hass' avocado trees. It is clear that bee and flower correlation is negative: The maximum flower density occurs during April, and the main bee activity starts only at the beginning of May. Early fruits appeared only when bee visitation started. This negative "r" means that the avocado flowers' relative attractiveness for the honeybees during its most blooming period was low, and the bees attracted to another bloom. In this case they attracted to Citrus and wild Mustard flowers, where they presented a significant positive correlation with flower density.
However, at May beginning the competing blooms almost ended and then the avocado relative attractiveness for the bees increased significantly.