code:

#special packages I utilized
from scipy.stats import gaussian_kde
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

trialdata = KDE_trial1csv    #I was lazy and opened the csv with all the data through menus, then named it "trialdata". This is just the excel sheet you sent me.

def getbins(values):
    kde = gaussian_kde(values)    #imposes a kernel density estimation on the data you feed into it (intended to be the "AVERAGE" column eventually)
    lower = min(values)-1
    upper = max(values)+1
    x_grid = np.linspace(lower, upper, 1000)        #populates a thousand evenly-spaced "grid points" along the x-axis for me to use in order to find the local minimums easily
    kde.set_bandwidth(0.035)       #sets the bandwidth factor for the KDE. 0.035 looks good but you can set this to whatever
    a = kde.evaluate(x_grid)           #uses the KDE to find the estimated density at each of the thousand points along the x-axis
    print plt.plot(a)                             #prints a plot so I can see the peaks and valleys
    mins = np.r_[True, a[1:] < a[:-1]] & np.r_[a[:-1] < a[1:], True]        #finds the density values (i.e. on the y-axis) that are the local minimums (i.e. smaller than the points on either side of it)
    cutoff_array = np.where(mins)            #finds the positions within the "x_grid" where those minimum values occurred (e.g. "the 54th point, the 134th point, the 189th point" etc)
    cutoffs = [(float(i)*(upper-lower)/1000+lower) for i in cutoff_array[0].tolist()]          #Translating the "x_grid positions" from the previous line back into the original scale to get cutoff points
    binlabels = pd.cut(trialdata["AVERAGE"], cutoffs, labels=range(len(cutoffs)-1,0,-1))      #use those cutoff points to make bins, labeled in reverse order so that the highest valued bin is tier 1.
    return binlabels
    
trialdata["Tier"] = getbins(trialdata["AVERAGE"])         #runs the script to make a new "Tier" column using data from the "AVERAGE" column.