-
Notifications
You must be signed in to change notification settings - Fork 7
/
T-GUI-Data-Viz-TS-Tools.qmd
281 lines (143 loc) · 17.6 KB
/
T-GUI-Data-Viz-TS-Tools.qmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
# GUI: data visualization and time series tools {.unnumbered #t-GUI-tstools}
## In this chapter
This chapter describes time series generic tools available in the Graphical User Interface:
<!-- (independent from a special routine SA, Bench...) -->
- data visualization
- spectral analysis tools
- aggregation
- differencing
- tests
Additional chapters related to GUI features, provide information on:
- [Overview](#t-gui-overview)
- [Specific Seasonal Adjustement and Modelling features](#t-gui-sa)
- [Output: series, parameters and diagnostics](#t-gui-output)
## Data visualization
*Container* includes basic tools to display the data. The following items are available: *Chart*, *Grid*, *Growth Chart* and *List*.
![** *Container* menu**](All_images/Container tools.jpg)
Several containers can be opened at the same time. Each of them may include multiple time series.
*Chart* plots the time series as a graph. This function opens an empty window. To display a given series drag and drop the series from the *Providers* window into the empty window. More than one series can be displayed on one graph. The chart is automatically rescaled after adding a new series.
![**Launching the *Chart* functionality**](All_images/A_Ref26.jpg)
The series to be viewed can be also dragged from the other windows (e.g. from the [*Variables*](#usr-def-var) window) or directly from the windows that display the results of the estimation procedure.
![**Displaying the seasonally adjusted series on a separate chart**](All_images/A_Ref27.jpg)
To adjust the view of the chart and save it to a given location use the local menu, which is displayed after right-clicking on the chart. The explanation of the functions available for the local menu is given below.
![**Local menu basic options for the time series graph**](All_images/A_Ref28.jpg)
To display the time series value at a given date, hover over it with the cursor. Once the time series is selected by clicking on it with the right mouse button, the options dedicated to this series are available.
![**Local menu options for chart**](All_images/A_Ref29.jpg)
A list of possible actions includes:
- **Open** -- opens selected time series in a new window that contains *Chart* and *Grid* panels.
- **Open with** -- opens the time series in a separate window according to the user choice (*Chart* *& grid* or *Simple chart*). The *All ts views* option is not currently available.
- **Save** -- saves the marked series in a spreadsheet file or in a text file.
- **Rename** -- enables the user to change the time series name.
- **Freeze** -- disables modifications of the chart.
- **Copy** -- copies the series and allows it to be pasted to another application e.g. into Excel.
- **Paste** -- pastes the time series previously marked.
- **Split into yearly components** -- opens a window that presents the analysed series data split by year. This chart is useful to investigate the differences in time series values caused by the seasonal factors as it gives some information on the existence and size of the deterministic and stochastic seasonality in data.
- **Remove** -- removes a time series from the chart.
- **Select all** -- selects all the time series presented in the graph.
- **Show title** -- option is not currently available.
- **Show legend** -- displays the names of all the time series presented on the graph.
- **Edit format** -- enables the user to change the data format.
- **Color scheme** -- allows the colour scheme used in the graph to be changed.
- **Lines thickness** -- allows the user to choose between thin and thick lines to be used for a graph.
- **Clear** -- removes all the time series from the chart.
- **Show all** -- this option is not currently available.
- **Export image to** -- allows the graph to be sent to the printer and saved in the clipboard or as a file in a jpg format.
- **Configure** -- enables the user to customize the chart and series display.
*Grid* enables the user to display the selected time series as a table. This function opens an empty window. To display a given series drag and drop the series from the *Providers* window into the empty window. More than one series can be displayed in one table.
![**Launching the *Grid* functionality**](All_images/A_Ref30.jpg)
To display options available for a given time series, left click on any time series' observation.
![**Local menu options for the *Grid* view**](All_images/A_Ref31.jpg)
The options available in *Grid* are:
- **Transpose** -- changes the orientation of the table from horizontal to vertical.
- **Reverse chronology** -- displays the series from the last to the first observation.
- **Single time series** -- removes from the table all time series apart from the selected one.
- **Use color scheme** -- allows the series to be displayed in colour.
- **Show bars** -- presents values in a table as horizontal bars.
- **Show crosshair** -- highlights an active cell.
- **Zoom** -- option for modifying the chart size.
When none of the series is selected, the local menu offers a reduced list of options. The explanation of the other options can be found below in the '*Local menu options for chart*' figure in the [*Container*](#container) section.
![**A reduced list of options for *Grid*** ](All_images/A_Ref32.jpg)
The *Growth chart* tab opens an empty window. Once a given series is dropped into it, *Growth chart* presents the year-over-year or period-over-period growth rates for the selected time series. More than one series can be displayed in a table. The growth chart is automatically rescaled after adding a new series.
![**The *Growth chart* view with a local menu**](All_images/A_Ref33.jpg)
A left click displays a local menu with the available options. Those that are characteristic for the *Growth chart* are:
- **Kind** -- displays m/m (or q/q) and y/y growth rates for all time series in the chart (previous period and previous year options respectively). By default, the period-over-period growth rates are shown.
- **Edit last year** -- for clarity and readability purposes, only five of the last years of observations are shown by default. This setting can be adjusted in the [*Options*](#options) section, if required.
The explanation of other options can be found below in the '*Local menu options for chart*' figure in the [*Container*](#container) section.
The *List* tab provides basic information about the chosen time series, such as; the start and end date, the number of observations and a sketch of the data graph. This function opens an empty window. To display information, drag and drop the series from the [*Providers*](#providers-window) window into the *List* window. A right click displays the local menu with all available options. Apart from the standard options, the local menu for *List* enables marking the series that match the selected frequency (yearly, half-yearly, quarterly, monthly) by using the *Select by frequency* option. An explanation of other options can be found below in the '*Local menu options for chart*' figure in the [*Container*](#container) section.
![**A view of a list of series**](All_images/A_Ref34.jpg)
For a selected series a local menu offers an extended list of options. The explanation of the functions available for the local menu is given below in the '*Local menu options for chart*' figure in the [*Container*](#container) section.
![**Options available for a selected series from the list**](All_images/A_Ref35.jpg)
## Spectral Analysis {#t-GUI-tstools-spec-graphs}
Spectral graphs are available from: *Tools* → *Spectral analysis.
![**Tools for spectral analysis**](All_images/A_Ref36.jpg)
![Tools for spectral analysis](All_images/image1_342.jpeg)
### Auto-regressive spectrum
When the first option is chosen JDemetra+ displays an empty *Auto-regressive spectrum* window. To start an analysis drag a single time series from the *Providers* window and drop it into the *Drop data here* area.
![Launching an auto-regressive spectrum](All_images/image2_342.jpeg)
When displaying an [Auto-regressive spectrum](#m-spectrum-ar-est) the number of observations, data transformations and other options such as the specification of the frequency grid and the order of the autoregressive polynomial (30 by default) can be specified by opening the *Window* → *Properties* from the main menu.
![**Auto-regressive spectrum's properties**](All_images/image3_342.jpeg)
The *Auto-regressive-Properties* window contains the following options:
- **Log** - log transformation of a time series;
- **Differencing**-transforms a data by calculating a regular (order 1,2..) or seasonal (order 4, 12, depending on the time series frequency) differences;
- **Differencing lag**-the number of lags that the program will use to take differences. For example, if *Differencing lag=3* then the differencing filter does not apply to the first lag (default) but to the third lag.
- **Last years**-a number of years at the end of the time series taken to produce autoregressive spectrum. By default, it is 0, which means that the whole time series is considered.
- **Auto-regressive polynomial order**-the number of lags in the AR model that is used to estimate the spectral density. By default, the order of the autoregressive polynomial is set to 30 lags.
- **Resolution**-the value 1 plots the spectral density estimate for the frequencies $\omega_{j}=\frac{2\pi j}{n}$, where $n \in (-\pi;\pi)$ is the size of the sample used to estimate the AR model. Increasing this value, which is set to 5 by default, will increase the precision of this grid.
### Periodogram
Choose *Tools* →*Spectral analysis* → *Periodogram* and drag and drop a series from the *Providers* window to the empty *Periodogram* window.
![Launching a periodogram](All_images/image5_342.jpeg)
The sample size and data transformations can be specified by opening the *Window* → *Properties*, in the main menu. The *Periodogram- Properties* window contains the following options:
- **Log** - log transformation of a time series;
- **Differencing**-transforms the data by calculating regular (order 1,2..) or seasonal (order 4, 12, depending on the time series frequency) differences;
- **Differencing lag**-the number of lags that you will use to take differences. For example, if *Differencing lag=3* then the differencing filter does not apply to the first lag (default) but to the third lag.
- **Last years**-the number of years at the end of the time series taken to produce periodogram. By default it is 0, which means that the whole time series is considered.
![An example of a periodogram](All_images/image7_342.jpeg)
![Periodogram's properties](All_images/image6_342.jpeg)
### Tukey spectrum
Choose *Tools* → *Spectral analysis* → *Tukey spectrum* and drag and drop a single series from the *Providers* window to the empty *Periodogram* window*.*
![Launching a Tukey spectrum](All_images//image8_342.jpeg)
The Tukey spectrum estimates the spectral density by smoothing the periodogram.
![An example of a Tukey spectrum](All_images/image9_342.jpeg)
Options for the Tuckey window can be specified by opening the *Window* → *Properties* from the main menu. The *Periodogram- Properties* window contains the following options:
- **Log** - log transformation of a time series.
- **Differencing** -transforms the data by calculating regular (order 1, 2..) or seasonal (order 4, 12, depending on the time series frequency) differences.
- **Differencing lag**-the number of lags that you will use to take differences. For example, if *Differencing lag=3* then the differencing filter does not apply to the first lag (default) but to the third lag.
- **Taper part**--parameter larger than 0 and smaller or equal to one that shapes the curvature of the smoothing function that is applied to the auto-covariance function.
- **Window length**--the size of the window that is used to smooth the auto-covariance function. A value of zero includes the whole series.
- **Window type**--it refers to the weighting scheme that it is used to smooth the auto-covariance function. The available windows types (*Square*, *Welch*, *Tukey*, *Barlett*, *Hamming*, *Parzen*) are suitable to estimate the spectral density.
![Tukey spectrum's properties](All_images/image10_342.jpeg)
## Aggregation {#aggregation}
*Aggregation* calculates the sum of the selected series and provides basic information about the selected time series, including the start and end date, the number of observations and a sketch of the data graph, in the same way as in the *List* functionality. *Aggregation* opens an empty window. To sum the selected series, drag and drop them from the *Providers* window into the *Aggregation* window. Right click displays the local menu with the available options. The content of the local menu depends on the panel chosen (the panel on the left that contains the list of the series and the panel on the right that presents the graph of an aggregate). The local menu for the list of series offers the option *Select by frequency*, which marks all the series on the list that are yearly, half-yearly, quarterly or monthly (depending on the user's choice). The explanation of the other options can be found below in the '*Local menu options for chart*' figure in the [*Container*](#container) section. The local menu for the panel on the left offers functionalities that are analogous to the ones that are available for the *List* functionalities, while the options available for the local menu in the panel on the left are the same as the ones available in *Chart* (see [*Container*](#container)).
![**The *Aggregation* tool**](All_images/A_Ref37.jpg)
## Differencing {#differencing}
The *Differencing* window displays the first regular differences for the selected time series together with the corresponding periodogram and the PACF function. By default, the window presents the results for non-seasonally and seasonally differenced series (( $d = 1,D = 1$)). These settings can be changed through the *Properties* window (*Tools* *→ Properties*). A description of a periodogram and the PACF function can be found [here](#periodogram).
![**The properties of the *Differencing* tool**](All_images/A_Ref38.jpg)
Typical results are shown below. The bottom left graph presents the partial autocorrelation coefficients (vertical bars) and the confidence intervals. The right-click local menu offers several functionalities for a differenced series. An explanation of the available options can be found below in the "*Local menu options for chart*" figure in the [*Container*](#container) section.
![The *Differencing tool](All_images/A_Ref39.jpg)
For the *Partial autocorrelation* and the *Periodogram* panels the right-button menu offers "a copy series" option that allows data to be exported to another application and a graph to be printed and saved to a clipboard or as a .jpg file.
## Tests {#t-GUI-tstools-tests}
Here we describe describe the GUI access to and display of tests. The underlying methods are detailed in [this chapter](#m-tests)
### Seasonality Tests
#### QS test
The test can be applied directly to any series by selecting the option *Statistical Methods \>\> Seasonal Adjustment \>\> Tools \>\> Seasonality Tests*. This is an example of how results are displayed for the case of a monthly series:
![qs](All_images/qs.png)
It is also visible in Main results panel and in diagnostics node.
#### Friedman test for stable seasonality
The test can be applied directly to any series by selecting the option *Statistical Methods \>\> Seasonal Adjustment \>\> Tools \>\> Seasonality Tests*. This is an example of how results are displayed for the case of a monthly series:
![friedman](All_images/friedman.png)
If the null hypothesis of no stable seasonality is rejected at the 1% significance level, then the series is considered to be seasonal and the outcome of the test is displayed in green.
It is also visible in Main results panel and in diagnostics node. (to be checked)
#### Identification of spectral peaks
##### In a Tukey spectrum {#t-GUI-tstools-id-p-tuckey}
The test can be applied directly to any series by selecting the option *Statistical Methods \>\> Seasonal Adjustment \>\> Tools \>\> Seasonality Tests*. This is an example of how results are displayed for the case of a monthly series:
![tktest](All_images/spectrum.png)
JDemetra+ considers critical values for $\alpha=1\%$ (code "T") and $\alpha=5\%$ (code "t") at each one of the seasonal frequencies represented in the table below, e.g. frequencies $\frac{\pi}{6}, \frac{\pi}{3}, \frac{\pi}{2}, \frac{2\pi}{3}\text{and} \frac{5\pi}{6}$ corresponding to 1, 2, 3, 4, 5 and 6 cycles per year in this example, since we are dealing with monthly data.
##### In AR Spectrum definition {#t-GUI-tstools-id-p-ar}
The test can be applied directly to any series by selecting the option *Statistical Methods \>\> Seasonal Adjustment \>\> Tools \>\> Seasonality Tests*. This is an example of how results are displayed for the case of a monthly series:
![artest](All_images/spectrum.png)
JDemetra+ considers critical values for $\alpha=1\%$ (code "A") and $\alpha=5\%$ (code "a") at each one of the seasonal frequencies represented in the table below, e.g. frequencies $\frac{\pi}{6}, \frac{\pi}{3}, \frac{\pi}{2}, \frac{2\pi}{3}\text{ and } \frac{5\pi}{6}$ corresponding to 1, 2, 3, 4, 5 and 6 cycles per year in this example, since we are dealing with monthly data.
##### In a Periodogram {#t-GUI-tstools-id-p-per}
The test can be applied directly to any series by selecting the option *Statistical Methods \>\> Seasonal Adjustment \>\> Tools \>\> Seasonality Tests*.
![periodtest](All_images/periodogram.png)
### Tests on residuals
Up coming content.