Zoomlar Projects
What is a Project?
Zoomlar was designed to ensure that you can generate accurate and relevant results that are specific to your data, but also so you can quickly and easily perform analysis as your data changes. This is facilitated in Zoomlar with “projects”
A “project” in Zoomlar is an analysis (either a “Trend” or “Predict” analysis) that is run on a set of data. The first time a project is created, Zoomlar will automatically build and tune the parameters of the analytic model, and then generate the results of the analysis. The project can optionally be saved and re-used at a later time on the same or new data. When a project is re-used there is no need for Zoomlar to build the analytical model and so results can be obtained very quickly.
What types of projects does Zoomlar support?
Zoomlar supports the “trend” and “predict” projects.
What is a trend project?
A trend project is one that uses time-based historical data and you are looking for time-based forecasts.
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | |
| Row1 | 12 | 131 | 45 | 67 | 8 | 10 | 11 | 12 | 14 | 15 | 16 | ? |
| Row2 | 21 | 32 | 76 | 87 | 29 | 87 | 67 | 7 | 32 | 24 | 23 | ? |
| Row3 | 34 | 32 | 87 | 54 | 43 | 86 | 65 | 8 | 24 | 54 | 8 | ? |
| Row4 | 54 | 56 | 65 | 67 | 45 | 54 | 46 | 98 | 53 | 56 | 89 | ? |
| Row5 | 67 | 67 | 78 | 53 | 65 | 32 | 33 | 6 | 56 | 7 | 8 | ? |
| Row6 | 78 | 65 | 45 | 12 | 67 | 56 | 67 | 46 | 78 | 7 | 8 | ? |
| Example 1. | ||||||||||||
The data can be in spreadsheet form (xls, xlsx and csv).
What time intervals can be used for trend projects?
A trend project can use any time interval. Whatever time period is used in the historical data (e.g. seconds, minutes, hours, days, months, years etc) will automatically be used for the forecast. You will also be given the option to specify specific times or ranges for the forecast.
The time intervals used in the historical data must be consistent, for example all months or all days. You can’t mix months with, for example weeks and hours.
How many time intervals can I forecast using a trend project?
This will depend on your historical data. Remember the further in to the future you try to forecast the more inaccurate the outliers will be. Zoomlar will advise you if the historical data supports multiple time intervals and if you are attempting to forecast too many time intervals.
How do I set the time intervals for a trend project?
Time intervals can be set within the Time period(s) input. These can be represented in the following ways by using the data from Example 1 above
| Time Period | Outputs |
| 1 | Dec |
| 1,3 | Dec, Feb |
| 1-3 | Dec, Jan, Feb |
What are examples of trend projects?
Examples of trend projects may include forecasting the following:
- Trend growth in sales of products on a per-shop or per sales person basis
- The utilisation levels of IT resources such as server CPU and storage data centre
- The consumption of items in stock for manufacturing process
What is a predict project?
A predict project is one that uses data that has a set of inter-dependent variables and you are looking to predict variables (?) in an incomplete dataset.
| Col1 | Col2 | Col3 | Col4 | Col5 | Col6 | Col7 | Col8 | Col9 | Col10 | |
| Row1 | 12 | 131 | 45 | 67 | 8 | 10 | 11 | 12 | 14 | 15 |
| Row2 | 21 | 32 | 76 | 87 | 29 | 87 | 67 | ? | ? | ? |
| Row3 | 34 | 32 | 87 | 54 | 43 | 86 | 65 | 8 | 24 | 54 |
| Row4 | 54 | 56 | 65 | 67 | 45 | 54 | 46 | ? | ? | ? |
| Row5 | 67 | 67 | 78 | 53 | 65 | 32 | 33 | ? | ? | ? |
| Row6 | 78 | 65 | 45 | 12 | 67 | 56 | 67 | 46 | 78 | 7 |
| Example 2. | ||||||||||
Are time intervals used for predict projects?
No, there is no need to use time-series data for predict projects.
What are examples of predict projects?
Examples of predicts projects may include predicting
- Analysing customer purchase history and demographics to predict likelyhood of purchasing a particular product
- Predicting manufacturing defects in a product line based on QA measurements
- Predicting how users of Cloud services will consume CPU cycles, RAM and storage space
Can I save projects for future use?
Yes. The first time a project is created, Zoomlar will automatically build and tune the parameters of the analytic model, and then generate the results of the analysis. The project can optionally be saved and re-used at a later time on the same or a new set of data. When a project is re-used there is no need for Zoomlar to re-build the analytical model, and so results can be obtained very quickly.
When should I save a project for future use?
Zoomlar was designed to ensure that you can generate accurate and relevant results that are specific to your data, but also so you can quickly and easily perform analysis as your data changes.
When you run an analysis with Zoomlar it develops a model and produces results. This model is based on the data you presented it with. Depending on the complexity and volume of data the process of developing the model may take some time. If you optionally save the project you are saving this model for future use. If you update your original data you can then re-use the saved project (i.e. the saved model) on the updated data.
In Zoomlar the default operation is for the project to be saved each time a new project is created. We recommend you use the default settings as the project only takes a small amount of disk space.
When can I re-use an existing trend project?
You can re-use a trend project if you satisfy the following:
- You want to re-run the analysis with the same historical data as used to create the project,
- If you make changes to the values in the historical data that was used to create the project,
- If you add rows to the historical data that was used to create the project,
- If you use an equal or greater number of columns than the window-size used to create the project,
- Any columns of text or strings in the new data are in the same columns as they were in the historical data used to create the project
- You want to forecast the same time-intervals.
When can I re-use an existing predict project?
You can re-use a predict project if you satisfy the following:
- You want to re-run the analysis with the same input data as used to create the project
- You make changes to the values in the input data that was used to create the project
- You have an equal or greater number of columns (input) as used to create the project
- Any columns of text or strings in the new data are in the same columns as they were in the historical data used to create the project
- You want to predict the same number of outputs