University rankings. the dataset on americancollege and university rankings (available from) contains information on 1302american colleges and universities offering anundergraduate program. for each university there are 17measurements, including continuous measurements (suchas tuition and graduation rate) and categoricalmeasurements (such as location by state and whether it is aprivate or public school).note that many records are missing some measurements. our first goal is to estimate these missing values from"similar" records. this will be done by clustering thecomplete records and then finding the closest cluster foreach of the partial records. the missing values will beimputed from the information in that cluster. a. remove all records with missing measurements fromthe dataset (by creating a new worksheet).b. for all the continuous measurements, run hierarchicalclustering using complete linkage and euclidean distance. make sure to normalize the measurements. examine thedendrogram: how many clusters seem reasonable fordescribing these data? c. compare the summary statistics for each cluster anddescribe each cluster in this context (e. g., "universitieswith high tuition, low acceptance rate. . "). hint: toobtain cluster statistics for hierarchical clustering, useexcel’s pivot table on the predicted clusters sheet. d. use the categorical measurements that were not used inthe analysis (state and private/public) to characterize thedifferent clusters. is there any relationship between theclusters and the categorical information? e. can you think of other external information thatexplains the contents of some or all of these clusters? f. consider tufts university, which is missing someinformation. compute the euclidean distance of this recordfrom each of the clusters that you found above (using onlythe measurements that you have). which cluster is itclosest to? impute the missing values for tufts by takingthe average of the cluster on those measurements.