Hey @rahajawadi! The different colors help give a quick visual as to what level of significance a factor is. For example, very dark blue means it's significant at the <0.00001 level, while a lighter blue may indiciate it's significant at the 0.01 level.    You are correct in that the factors with no bars have been eliminted using backwards selection. You can confirm this by expanding the object in the top-right corner, then selecting the "Selection Summary" tab:     All analytics, statistical, and machine learning objects in Visual Analytics have more information to show if you hit the expand button. Some even come with natural language generation. Try out some other models and see what you can find! ... View more

Hey @yac431! @PaigeMiller is correct. For this particular object, aggregations do not affect it. The correlations are derived on un-summed data based on the Pearson Product Moment Correlation Coefficient. This formula is:   ... View more

Hey @Sigord! From your description, it sounds like you're doing this: Drop-down filter with a year in data source 1 Year parameter in drop-down filter, 'year'p Advanced data source filter in data source 2: 'year'n = 'year'p Can you confirm if the parameter is being set to a value when the report is opened? If not, this would cause 'year'p to initially not start off with a value.   Since both data sources share Year, have you tried mapping Year from data source 1 to Year in data source 2? That way you can avoid a parameter-based solution. ... View more

Hey @znhnm! The day of the week is what's causing your issue. anydtdte cannot parse that particular pattern, but it will parse the string if you remove the day of the week. You can fix this by removing the first word: parse(removeword('CatDateTime'n, 1), 'anydtdte32.')   Note that anydtdte returns a date while anydtdtm returns a datetime. If you return a datetime, you can still format it as a month and retain the original datetime data by duplicating the column.   ... View more

Hey @illmatic! Based on the header of your list table, if your expression is this: abs('forecast'n - 'actuals'n) Then you should be good.   Based on what you have, it looks like Total is a part of your list table, which is why it's aggregating that way; however, this is the correct calculation for your overall forecast. Consider the following hypothetical forecast where the Total row is the sum of each column:   Product Forecast Actual Error Abs Error A 0 100 -100 100 B 300 200 100 100 C 300 300 0 0 Total 600 600 0 200   Total is a bottom-up forecast created by summing up products A, B, and C. The individual forecasts of A and B are off by -100 and 100 respectively, but the overall bottom-up forecast has an error of 0: the errors in products A and B cancel each other out to create a perfect overall forecast. If you're trying to figure out how good your individual forecasts are, this can be a deceptive metric if your individual forecasts aren't great. The sum of the absolute errors tells us there are 200 total errors among all of the individual forecasts, but it does not mean that the overall forecast has an absolute error of 200.    If you want to capture the total number of absolute errors of all forecasts, you need to remove the Total row and have Visual Analytics calculate the sum for you with the Totals option in a list table:   You can filter out your Total row with an object filter so you do not need to reload the data:   However, summing up all of the absolute errors is a relative measure and doesn't necessarily tell the whole story. One forecast could have a lot of errors while other forecasts could have a few. It's one indicator that there may be a problem with one or more forecasts, but it cannot be used to judge the accuracy of all hierarchical forecasts, especially if one hierarchy has large values while another hierarchy has low values. If you would like some good tips on how to judge the average accuracy of your hierarchical forecasts, I would recommend posting in the Forecasting and Econometrics forum.   Generally I have always looked at MASE or MAPE values and sorted them in ascending order to view how well individual forecasts are doing. ... View more

Hey @Kritiksharma123! Migrating a SAS 9.4 report to SAS Viya is very easy. Steps 3-8 below also apply if you have a JSON file: Load all of your necessary datasets to a CASLIB in SAS Viya. Export your report(s) as a SAS package (.spk file) from SAS Management Console. Be sure "Include dependent objects when retrieving initial collection of objects" is selected. For more information on exporting SAS packages, see this link. Log in to SAS Viya. In the hamburger menu, navigate to "Administration --> Manage Environment"   Select the Import tab: Upload your .spk file and follow the on-screen instructions. Map your report's data sources to the data sources you loaded to the CASLIB in step 1. Done! Your report should now be available and ready to go. ... View more

You can either select it when exporting the data, or you can enable it within the list table itself. On the right-hand Options pane, scroll down Table and select the checkbox for "Detail data."   Exporting:   Or list Table options:   ... View more

Hey @DaveShea! You cannot put carriage returns within titles, but you can use a text box above your visualization to create a title and add a description. You can reduce the padding of each object to make the text a little closer to the chart. Unfortunately when exporting the bar chart as an image the text will not be there, but a simple screen grab or snipping tool can solve that. ... View more

PROC SEVERITY will estimate the distribution of data and give you the closest estimates. You can choose the distributions you'd like it to estimate and it will identify the most likely distribution and its estimates. Since you know it's log normal, you can tell it to estimate only the parameters of a log normal distribution.   Let's simulate a lognormal distribution with Mu=0 and Sigma=0.5: data logn; call streaminit(12345); do i = 1 to 10000; x = rand('lognormal', 0, 0.5); output; end; run;     Next, let's use PROC SEVERITY to estimate the parameters of this distribution assuming we know it is log normal. proc severity data=logn; loss x; dist logn; run;   While not perfect, it's pretty close at getting the right answer: Mu is very close to 0, and Sigma is within about 0.004 decimal points of the true value.    Let's say you didn't know the distribution. You can test a variety of distributions, including custom distributions, with PROC SEVERITY.  proc severity data=logn; loss x; dist _ALL_; run; Even after testing all of these distributions, it still is able to determine that x is most likely log normal. ... View more

Hey @nouramohammad4! It looks like you've created a new machine learning pipeline from Visual Analytics. To define partitions, you'll need to have a partition indicator. When you create a new project within Model Studio, you have the option of creating this partition indicator. When you create a new machine learning pipeline through Visual Analytics, you partition it within Visual Analytics. To partition it: In the Data Pane in Visual Analytics, click New Data Item and select Partition.    Partition your data into Training, Validation, and Test as you'd like. Optionally give it a random number seed and a unique ID to create deterministic partitions. This will make sure you have the same random sample if the data is reloaded or if you create a new report. For more information about setting a unique row identifier, see this documentation link.   Add the Partition data item to your model's Partition ID role. Create a new pipeline or add to an existing pipeline. You should now see a Partition role automatically added as a variable to your data, usually named something like: va__d__PARTITION<numbers>__ONES. Your pipeline will automatically use this partition. ... View more