Please authenticate






For any question, please contact administrator



If utilization >= 1, check the # servers selected.


Table 1. The graph below shows the distribution of time spent in the queue and the system.

Deciding on # of Staff Required


Below is shown the minimum number of servers required to meet customer demands, using a 99% confidence level (aka Grade of Service).

Waiting Probabilities

Below is shown how likely is that a customer will face a queue on arrival, and how this varies with different # of servers. If an inappropriate range is selected, the the probability might exceed 1.



Loading...

Loading...


Server Utilization and Free Time

The table shows the amount of free time across all servers per day given a flow of 400 customers. The average daily free time per server is also included. If we have a sample size of 4000, this would translate in 10 simulations of the day.

% Share of people waiting more than the selected threshold


Queue Proportions

This table shows the various proportions of the queue size

System Proportions

This table shows the various proportions of the system size, being people served + people in the queue.

Data Loader


                                      

Reading parameters:

DB Connections

Data Preview

Forecast Level Selection


Loading...

Data Filtering

  100 / 100
100%

Missingness Stats

Summary of Missing Data

Loading...

Feature Engineering


Loading...
Barplots of the variables recognized as categorical/factors.
Loading...

Figure Preview

Loading...
Loading...

Distribution Fitting



Loading...
Drag the mouse over the graph and double click to zoom-in. Restore by clicking twice on the graph afterwards.
Loading...
Loading...
Loading...
Loading...
Loading...



Loading...


Loading...


Series Importance to Portfolio

Observe what series gain or lose importance in your portfolio.

Loading...





Autocorrelations and Spectrograms

In general the ACF and PACFs are often used for time series order selection. The rules are outlined in the table below:

However, for more complex models as in our case we cannot use them for order selection and will have to compare models more rigorously based on their Akaike Information Criterion.

Loading...


ARIMA Model Specification

ETS Model Specification

TBATS Model Specification

THETA Model Specification

Prophet Model Specification

  • Table 0 provides a summary statistics of each time series level to be modelled covering a wide spectrum of useful information from basic summary statistics to stationarity tests and entropy.

Model Competition Rankings

Table 1. gives a complete overview of the model competition. The number of rows depends on the series modeled multiplied by the # models used in the model competition. For each combination the error metrics, e.g. Mean Absolute Percentage Error (MAPE) are provided. For more information on the error metrics, please refer to the documentation. The models are ranked by their MAPE across each series id. Additional information is provided for each model by clicking on the 'Model Info' button tab.

Loading...
  • Table 2. summarizes the results from the model competition by showing which models won the most competitions and what is their average performance across all series.

Forecasting Results Exploration

Model Testing Phase



Loading...




Final Forecast Results

Loading...

Level:



Loading...



Validation Phase

Train/Test Splits

Loading...

Model Performance

Loading...



Detailed Results

Loading...



Loading...

Loading...
Loading...

Loading...

See Graphs

Loading...
Loading...