The consumption spike detection and replacement algorithm is a machine learning-assisted algorithm that detects spikes from the input timeseries data and replaces the anomalous spike values. The algorithm works on both cumulative and interval type of data series. Underlying advanced machine learning algorithms and statistical methods segregates the spikes from the input data stream and computes the replacements, all happening in near-real time. Minimal human effort is expended for its operation as the algorithm intelligently adapts to your requirement to serve your various needs. The results of the detection and cleansing process are available to view on an intuitive dashboard and to download in the form of a spreadsheet.