applied_macroeconomics

Applied Macroeconomics, a course taught at the University of Warsaw

arima-for-demand-forecasting

This is the source code for the blog at https://rpubs.com/Tesfahun_Boshe/arimaforcasting which is aimed at looking at ARIMA models in-depth.

bootstrapping-for-the-best-regression

classification-censusincome-ml

classification-of-census-income

covid19_vaccination-shiny

epl-seasons-analysis

An automated reporting of EPL seasons per Team using RMarkdown.

erlangwithpython

europeanoptionpricing

how-to-convert-schedule-to-per-interval-coverage

Let’s say you have just created a schedule for your workforce. If your workforce has the same shift start and shift end in all days, it should be easy to find how many heads you will have per interval. What if that was not the case? Most WFM schedulers have to create much more dynamic schedules with not only multiple shift start and shift end but also with shifts going across days. It is also possible that agents do not work all the days, and the shifts are of varying length. Once you have created such schedule, how do you convert it to coverage data to check your efficiency per interval? I have seen some excel solutions for this but let’s see how we can do this in much easier way using R. At the end of this, we will plot the coverage against the predetermined requirement and calculate the schedule efficiency.

plotting-with-ggplot2

predicting-the-net-hourly-electrical-energy

resourceplannint2hiringplan-wfm

Currently there are not known algorithms for converting per-interval FTE requirements to number of heads needed. This repo aims at introducing one. The associated Shiny App can be used for a quick conversion of your FTE's per interval to the most optimal schedule and number of heads required.

rrcourse2022

This is a course repository for Reproducible Research 2022

rwfm

stock_price_scraping

tesfahunboshe

webapp_with_flask

wfm-forecasting-with-r

Forecasting is the first important stage of workforce management planning. WFM forecastors create the forecast of ,including others, volume of conversations to be expected in some future time. Forecasting future contacts is a difficult task because of the uncertainty associated with the many factors that determine contacts as well as the nature of distribution of contacts in time. Time series data is an object observed in many consecutive units of time in chronological order. While there is not a single best method for forecasting, there are widely used methods and best practices. We will discuss some of the most commonly used time series models and hybrid approaches in this blog. We will also cover the metrics which can indicate the accuracy of the forecast. accuracy() function from forecast package gives mean error (ME), root mean square error (RMSE), mean absolute error (MAE), mean percentage error (MPE), mean absolute percentage (MAPE), mean absolute squared error (MASE), autocorrelation function index (ACFI) and Theil’s U values. MAPE is considered the best accuracy measure since it is not sensitive to sign of error and the magnitude of units. We will use MAPE measures to compare the accuracy of different methods to be discussed. The html output is currently available at rPubs: https://rpubs.com/Tesfahun_Boshe/873021

wfm-planning-with-r

workforce-management-app-with-shiny

This app was developed using Shiny package. It can be used for workforce management purposes for small and medium size call centers of similar purposes. The app has a page for schedule view, another for the coverage view per market and per channel, and another two pages for submitting preferences and viewing team schedule for the selected day range.