## Rows: 12,299
## Columns: 17
## $ task_number <chr> "1735", "1742", "1971", "2134", "2251", "2283",~
## $ summary <chr> "Flag RI on SCM Message Summary screen using me~
## $ priority <dbl> 1, 1, 2, 5, 10, 1, 5, 5, 6, 5, 2, 1, 3, 1, 1, 8~
## $ raised_by_id <chr> "58", "58", "7", "50", "46", "13", "13", "13", ~
## $ assigned_to_id <chr> "58", "42", "58", "42", "13", "13", "13", "58",~
## $ authorised_by_id <chr> "6", "6", "6", "6", "6", "58", "6", "6", "6", "~
## $ status_code <chr> "FINISHED", "FINISHED", "FINISHED", "FINISHED",~
## $ project_code <chr> "PC2", "PC2", "PC2", "PC2", "PC2", "PC9", "PC2"~
## $ project_breakdown_code <chr> "PBC42", "PBC21", "PBC75", "PBC42", "PBC21", "P~
## $ category <chr> "Development", "Development", "Operational", "D~
## $ sub_category <chr> "Enhancement", "Enhancement", "In House Support~
## $ hours_estimate <dbl> 14.00, 7.00, 0.70, 0.70, 3.50, 7.00, 7.00, 7.00~
## $ hours_actual <dbl> 1.75, 7.00, 0.70, 0.70, 3.50, 7.00, 7.00, 7.00,~
## $ developer_id <chr> "58", "42", "58", "42", "13", "13", "43", "58",~
## $ developer_hours_actual <dbl> 1.75, 7.00, 0.70, 0.70, 3.50, 7.00, 7.00, 7.00,~
## $ task_performance <dbl> 12.25, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00~
## $ developer_performance <dbl> 12.25, 0.00, 0.00, 0.00, 0.00, 0.00, NA, 0.00, ~
## # A tibble: 1 x 6
## projetos categoprias sub_categorias estimativas tasks prioridades
## <int> <int> <int> <int> <int> <int>
## 1 20 3 24 12299 10266 10
Temos 20 projetos, com 12299 estimativas de tempo para tasks. Não há apenas uma estimativa por tarefa, já que há apenas 10266 valores distintos de task_number.
## # A tibble: 10,268 x 4
## task_number category summary n
## <chr> <chr> <chr> <int>
## 1 10605 Management Staff Meeting 8
## 2 6889 Management SiP Staff Meeting 8
## 3 10089 Operational Office Move and bits and bobs 7
## 4 10974 Management Extended SiP Lunch 7
## 5 11056 Management SiP Company Meeting 7
## 6 11270 Management Staff Meeting 7
## 7 13124 Management Company Meeting - scorecard and discussion 7
## 8 13190 Management Marketing management meeting 7
## 9 13253 Management YYY ZZZ's Marketing presentation and meeting 7
## 10 3812 Development Weekly Developer Meeting 14th September 2005 -~ 7
## # ... with 10,258 more rows
## # A tibble: 24 x 3
## sub_category tasks correlacao
## <chr> <int> <dbl>
## 1 Board Meeting 21 0.911
## 2 Business Specification 96 0.861
## 3 General Documentation 254 0.857
## 4 Support 1045 0.837
## 5 Client Support 616 0.829
## 6 Enhancement 2592 0.821
## 7 Technical Specification 61 0.807
## 8 Third Party 18 0.805
## 9 Conversion 75 0.799
## 10 Documentation 72 0.788
## # ... with 14 more rows
To provide examples of the relationship between the estimates and actual hours taken in the company as a whole and what this relationship looks like in different task subcategories. We present the correlation of the variables hours estimate and hours current in graph 1. The relationship of the two variables is shown in the log scale on base 10. In this graph, we have the hours estimate visually in the vertical axis and the current hours taken in the company in the horizontal axis. Given this, we can identify for example that the Board Meeting subcategory has a very strong correlation and Research has a weaker correlation as shown in the illustration shown in this graph.
In graph 2, we have the evidence of the relationship between the team size and the average abs error of the team estimates visually. According to these samples, there is a noticeable relationship between the size of the team and is related to the average error of the team’s estimates. Therefore, it is concluded that the graph indicates when there are atypical observations, outliers and extremes as it is possible to be studied in the graph.
In graph 3, it expresses the evidence of the relationship between task priority and the error in its estimate taken at the company. In this graph, the task’s priorities are organized on the horizontal axis and errors in its estimates are organized on the vertical axis of the graph. The results are presented on the square root scale. According to Kendall’s correlation of -0.02498766, it revealed no relationship. We can conclude that there is no positive or negative relationship between priority and absolute error, that is, the developers in these results, make similar mistakes in low-priority and high-priority estimates. It is concluded that the error is similar. We highlight only one outlier for priority tasks one and five, as it obtains an atypical value for the other tasks, as shown in the evidence in this graph.