Using Forecasts within Orixa 

Not all Apps need the facilities of forecasting, so it is not included in the basic framework, but is added using the standard Orixa mechanism: A data-table named "Forecasts" is added as a BusinessObject, and once present this triggers the addition of Forecasting related features in the App.

Using Forecasting is technically complex to understand to begin with, but once understood and implemented it can be used by ordinary users of an App without further intervention from the Developer.

How Forecasts Work

The Forecasts data-table follows a standard pattern in Orixa, using a "LinkTable" and "LinkID" field. This means that you can store values in the Forecasts data-table, and link them either to a whole data-table (via the LinkTable) or to a specific record (via the LinkID).

You add the new Forecasts BusinessObject using a SQL Script. This script is shown at the end of this page. Once it has been added you enable any other BusinessObject in your App to create linked Forecasts records by adding the DESCRIPTION AddForecasts=1 to its data-definition, as shown in the SQL below.

CREATE TABLE "Customers"
(
"ID" INTEGER NOT NULL,

[OTHER FIELDS AND CONSTRAINT DEFINITIONS WOULD FOLLOW HERE]

)
DESCRIPTION '[Properties]
AddForecasts=1' 

You can also ALTER a data-table to enable Forecasts:

ALTER TABLE "Customers" DESCRIPTION '[Properties]
AddForecasts=1'

Once this Description is added new features will be visible in Orixa.

Entity Insert Forecasts  

If the above change is made, on the System Entities Screen a new Action "Insert New Forecasts Linked to Customers" will be added, as shown at 1.

Adding Forecasts here sets the LinkTable of the Forecasts data, but leaves the LinkID blank. This allows data entered to be returned and compared to sum data for the LinkTable.

As an example, "Monthly Sales Forecasts" could be added, and compared to the sum of values in a Sales data-table.

Record Insert Forecasts  

Also, on the individual Records a new Action is created, as shown at 1.

Adding Forecasts here sets both the LinkTable and LinkID for the Forecasts record. This allows data to be queried and returned per customer. Or per any other data-record in any data-table (Products, Projects, People etc.)

Both the above actions open a Forecasts data-entry Grid (shown below), which allows the user to add multiple records to the Forecasts data-table.

Forecasts batch-entry grid  

  1. Give your set of Forecasts a name. This will allow you to select them in SQL scripts later.
  2. Set a start date for the first Forecast.
  3. Choose whether the Forecast should be Weekly, Monthly or Annual.
  4. Enter the number of records, records will be added with a start-date separated by the number of days determined by the chosen "PeriodType".
  5. Enter Quantity and Value amounts for the start of the period and end of the period. 

Forecasts batch-entry grid 02  

  1. Once you have completed setting the values and range for your Forecasts, click Activate and a data-grid will open showing the added records.
  2. Note that the values across the period will be automatically computed to vary between your start and end value.
  3. Once you are happy with the values in the Grid, cllick Insert New Forecasts and data will be added to the Forecasts data-table.

Note: Once the Grid is activated, it is not fixed. You can edit the values to change them to meet your needs. For example you may plan for Sales to grow over the year, but you may expect one month to be below average, due to seasonal issues or an annual holiday. In such cases you can manually correct individual values of the data before you click "Insert"

Adding Individual Forecasts Records

You may not need to add large numbers of Forecasts records at the same time. For example you might just want to set a "Sales Target" for a Customer or Sales-person.

In this case you do not need to use the Insert Forecasts Action. Instead, just add individual Forecasts records.

Forecasts single record entry  

From any data-record in a BusinessObject that has Forecasts enabled, click the "New" button, then enter Forecasts data, as Shown at 1.

Forecasts Edit Form  

The Forecasts Edit Form, shown on the left, contains fields which can be used to set values and quantities over time, linked to a Table or Record.

Users (provided they have adequate security) can add data in these forms.

Note that the Forecasts Edit Window has a built in Action "Duplicate this record", allowing users to easily create multiple records in quick succession.

 

Using Forecasts data in a Dashboard, Resource, Cube or Chart

Entering Forecasts is simple. Writing the SQL Scripts which will create visualizations of your Forecasts is a bit more technical. Usually the SQL will be writen by an Orixa Developer and added as a record to the "Resources" BusinessObject. Once the SQL is complete, any user can open the dashboard and see the result in a location in your App set by fields in the Resources data-record.

Creating a Forecast Resource  

In the above example, an App includes features to monitor Hourly Billing by staff. Details of this billing are stored in a WorkItems data-table. Forecasts have been added with LinkTable = WorkItems and Name = Paid Hours.

Note how the above SQL relies on sub-select statements. That is the data is first returned and summed for Forecasts and WorkItems, then this data is pulled together to create a result-set.

The Resource has been created with "ComponentID" = Chart, so a chart will be shown when the Resource is activated.

Accessing Forecast Dashboard  

The above image shows how the newly added Forecast Resource shows in the App: Accessible via the "Reports / Dashboards" for the "WorkItems" BusinessObject.

Forecast Dashboard  

The above image shows the resulting chart, with "Target" in green, "Actual" in blue, and over / under performance shown as a purple line, marked 1.

Adding forecasting features to your Orixa App involves several stages:

1. Run the SQL script needed to create the Forecasts BusinessObject in your App. Sample SQL doing this is added at the end of this article.

2. Link chosen BusinessObjects in your App to the Forecasting system by adding the correct DESCRIPTION to their data-definition.

3. Add Forecasts records with your target values and quantities, linked to specific BusinessObjects and data-records.

4. Write SQL to query the Orixa database linking the Forecasts to actual business data, and add this as a new "Resources" record.

5. Run the resource and view the resulting data. If the Resource is a Grid it will appear fully formed. If it is a Cube or Chart the user will need to design it to see the final result.
 

While setting up a Forecasts system is complex, once complete further work will be minimal. Users can add Forecasts for a period, and see the resulting data immediately in linked Resources.

Forecasts BusinessObject

It is important to remember that Forecasts are just an ordinary BusinessObject like any other BusinessObject in Orixa.
This means once Forecasts data has been added it can be viewed, edited, updated or deleted by users exactly like any other records in the system.
Linked Forecasts can also be seen in a grid directly from records they are connected to.

Forecasts Entity  

Once Forecasts have been added users (with the correct Security Level) can access a data Grid from which they can search, edit and update records.

It is possible to add Resources linked to Forecasts, so you can show specialized visualizations of Forecasts data if you want to.

View Linked Forecasts  

If a BusinessObject is linked to Forecasts, then an item is added to the list of Grids in that BusinessObject's Edit Window (shown at 1. on the left) allowing users to see linked Forecasts records, to allow updating.

Example SQL For Forecasts Resources

The following SQL will return data-sets which can be shown in your Orixa App bringing together Forecasts for a period with data in your App.

SELECT
  FC.YYMM,
  FC.Target,
  WI.Actual,
  ROUND((WI.Actual - FC.Target) / FC.Target * 100, 2) as "Percentage" 
FROM
  (SELECT
  YearMonth(CAST(DateStart as DATE)) as YYMM,
  SUM(Quantity) as Target
  FROM Forecasts
  WHERE LinkTable = 'WorkItems'
    AND Name = 'Paid Hours'
    AND DateStart >= CAST([MinDate] as DATE)
    AND DateStart <= CAST([MaxDate] as DATE)
  GROUP BY YYMM) as FC
LEFT JOIN
  (SELECT
  YearMonth(W.DateDone) as YYMM,
  SUM(W.HoursWorked) as Actual
  FROM WorkItems W
  LEFT JOIN ContractItems C ON C.ID = W.ContractItemsID
  WHERE C.Chargeable = true
  AND W.DateDone >= CAST([MinDate] as DATE)
    AND W.DateDone <= CAST([MaxDate] as DATE)
  GROUP BY YYMM) as WI ON WI.YYMM = FC.YYMM 

The above SQL generates a "Target" field, from a SUM of the Quantity field of Forecasts, and an "Actual" field from a SUM of the "HoursWorked" field of the WorkItems data-table.

Only Forecasts with a LinkTable = WorkItems and Name = Paid Hours are returned, so other Forecasts are not accidentally added to the totals.

Note how the SQL uses the YearMonth Function to create a field "YYMM", and how this Period-field is used to JOIN the two sets of data, so that Target and Actual for each YYMM appear in the same row of the result-set.

SELECT
  FC.YYMM,
  FC.Target,
  COALESCE(E.Actual, 0) as Actual,
  ROUND((E.Actual - FC.Target) / FC.Target * 100, 2) as "Percentage"
FROM
  (SELECT
  YearMonth(DateStart) as YYMM,
  SUM(Value) as Target
  FROM Forecasts
  WHERE LinkTable = 'Expenses'
  AND Name = 'Monthly Expenses'
  AND DateStart >= CAST([MinDate] as DATE)
  AND DateStart <= CAST([MaxDate] as DATE)
  GROUP BY YYMM) as FC
LEFT JOIN
  (SELECT
  YearMonth(DateDone) as YYMM,
  COALESCE(SUM(Value), 0.00) as Actual
  FROM Expenses
  WHERE DateDone >= CAST([MinDate] as DATE)
  AND DateDone <= CAST([MaxDate] as DATE)
  AND NOT ExpensesTypeID = 56675
  GROUP BY YYMM) as E ON E.YYMM = FC.YYMM 

The above SQL generates a "Target" field from a SUM of the Value field of the Forecasts table, and an "Actual" field from the SUM of Monthly expenses. In this case (as with the previous example) the result is totalled per month, but it would also be possible to GROUP BY other fields, such as by StaffID to show totals for each staff member.

SELECT
  FC."Year",
  W.Customer,
  FC.Target,
  COALESCE(W.Actual, 0) as Actual,
  COALESCE(W.Planned, 0) as Planned, 
  ROUND((W.Actual - FC.Target) / FC.Target * 100, 2) as "PercentageActual",
  ROUND((W.Planned - FC.Target) / FC.Target * 100, 2) as "PercentagePlanned"
FROM
  (SELECT
  EXTRACT(Year FROM DateStart) as "Year",
  LinkID,
  SUM(Value) as Target
  FROM Forecasts
  WHERE LinkTable = 'Customers'
  AND Name = 'Customer Annual'
  AND DateStart >= CAST([MinDate] as DATE)
  AND DateStart <= CAST([MaxDate] as DATE)
  GROUP BY "Year", LinkID) as FC
LEFT JOIN
  (SELECT
  EXTRACT(Year FROM DateDone) as "Year",
  O.ID as LinkID,
  O.FullName as Customer,
  COALESCE(SUM(WI.HoursPlanned * 60), 0.00) as Planned,
  COALESCE(SUM(Value), 0.00) as Actual
  FROM WorkItems WI
  LEFT JOIN Contracts C ON C.ID = WI.ContractsID
  LEFT JOIN Organisations O ON O.ID = C.CustomersID
  WHERE DateDone >= CAST([MinDate] as DATE)
  AND DateDone <= CAST([MaxDate] as DATE)
  GROUP BY Year, LinkID) as W ON ((W.Year = FC.Year)
                              AND (W.LinkID = FC.LinkID)) 

The above SQL generates 2 grouping Columns, and is intended to show results in a Cube, rather than a chart.

Two sub-select statements return columns for Year, LinkID and Target / Actual. The two grouping Columns are then used in the JOIN Statement. Note that this Forecast also uses the "HoursPlanned" and "Value" columns, allowing Forecasts to be compared with actual planned work which staff have entered into the App, but which has not yet been completed.

Decision Cube output from example SQL  

Decision Cube, toggled to show as a Chart  

Note that Decision-Cubes can also display chart data.

Forecasts data-table structure

Forecasts data-table structure  

The Forecasts data-table has the following structure:

CREATE TABLE "Forecasts"
(
"ID" INTEGER DEFAULT UID() NOT NULL,
"DateStart" DATE,
"Name" VARCHAR(120) COLLATE "ANSI" DESCRIPTION '[Properties]
ReuseLast=1
Distinct=1',
"PeriodTypeID" INTEGER DEFAULT TypeID('Monthly') DESCRIPTION '[Properties]
ReuseLast=1',
"Value" DECIMAL(19,4) DEFAULT 0 NOT NULL,
"Quantity" FLOAT DEFAULT 0 NOT NULL,
"LinkTable" VARCHAR(60) COLLATE "ANSI" NOT NULL,
"LinkID" INTEGER,
"StaffID" INTEGER DESCRIPTION '[Properties]
ReuseLast=1',
"DateCreated" TIMESTAMP DEFAULT Current_Timestamp,
"Complete" BOOLEAN DEFAULT false NOT NULL,
CONSTRAINT "PK_Forecasts" PRIMARY KEY ("ID"),
CONSTRAINT "StaffID" FOREIGN KEY ("StaffID") REFERENCES "Staff" ("ID")
  ON UPDATE NO ACTION ON DELETE NO ACTION,
CONSTRAINT "PeriodTypeID" FOREIGN KEY ("PeriodTypeID") REFERENCES "Types" ("ID")
  ON UPDATE NO ACTION ON DELETE NO ACTION
)
DESCRIPTION '[Properties]
AddDuplicateRecord=1'

 

Forecasts data-structure  

 

  1. The LinkTable and LinkField allow "promiscous linking" between any Forecasts data-record and other records or data-tables in your App.
  2. Each row contains two number columns: "Value" and "Quantity" these can be used to store the data you want to set as your Forecast for that period.

 

SQL Script to add Forecasts to your Orixa App

Note that the script below requires the following BusinessObjects to already be present in your Orixa App: People, Stafff.

-- ORIXA AUTOMATED SQL TO CREATE BUSINESS-OBJECT --
-- MAIN CREATE SCRIPT FOR Forecasts --
CREATE TABLE "Forecasts"
(
"ID" INTEGER DEFAULT UID() NOT NULL,
"DateStart" DATE,
"Name" VARCHAR(120) COLLATE "ANSI" DESCRIPTION '[Properties]
ReuseLast=1
Distinct=1',
"PeriodTypeID" INTEGER DEFAULT TypeID('Monthly') DESCRIPTION '[Properties]
ReuseLast=1',
"Value" DECIMAL(19,4) DEFAULT 0 NOT NULL,
"Quantity" FLOAT DEFAULT 0 NOT NULL,
"LinkTable" VARCHAR(60) COLLATE "ANSI" NOT NULL,
"LinkID" INTEGER,
"StaffID" INTEGER DESCRIPTION '[Properties]
ReuseLast=1',
"DateCreated" TIMESTAMP DEFAULT Current_Timestamp,
"Complete" BOOLEAN DEFAULT false NOT NULL,
CONSTRAINT "PK_Forecasts" PRIMARY KEY ("ID"),
CONSTRAINT "StaffID" FOREIGN KEY ("StaffID") REFERENCES "Staff" ("ID")
  ON UPDATE NO ACTION ON DELETE NO ACTION,
CONSTRAINT "PeriodTypeID" FOREIGN KEY ("PeriodTypeID") REFERENCES "Types" ("ID")
  ON UPDATE NO ACTION ON DELETE NO ACTION
)
DESCRIPTION '[Properties]
AddDuplicateRecord=1'
!
-- INSERT BUSINESS OBJECTS DATA RECORD Forecasts --
INSERT INTO "BusinessObjects"
(Name, Color, BusObjType,
AddNewButton, LinkToImages, LinkToComments, LinkToAddresses,
LinkToPhones, SecurityLevel, Icon,
DisplayScript, ListScript, ViewScript, DiaryScript,
SummaryScript, Settings)
VALUES
('Forecasts', 16777215, 'N', True, False, False, False, False, 0, 181,
  ' SELECT
    Name, ID
FROM "Forecasts"
WHERE ID = %d ', ' SELECT
  Name as FullName, ID
FROM "Forecasts"
ORDER BY FullName ', 'SELECT
  ID,
  DateStart,
  Name, 
  T1.Name as PeriodType,
  "Value",
  "Quantity",
  Pe.FullName as Staff,
  DateCreated,
  LinkTable,
  Complete
FROM "Forecasts" F
  LEFT JOIN Types T1 ON T1.ID = F.PeriodTypeID
  LEFT JOIN People Pe ON Pe.ID = F.StaffID
%s ',
  '', '', '')
!
-- INSERT SEARCHES DATA RECORDS --
INSERT INTO Searches
(LinkTable, Name, SQLStr)
VALUES
('Forecasts', 'All Records', '')!
-- INSERT SEARCHES DATA RECORDS --
INSERT INTO Searches
(LinkTable, Name, SQLStr)
VALUES
('Forecasts', 'Recently Added',
' WHERE DateCreated > Current_Timestamp - INTERVAL ''1'' MONTH ')!
-- INSERT SEARCHES DATA RECORDS --
INSERT INTO Searches
(LinkTable, Name, SQLStr)
VALUES
('Forecasts', 'Search "Name"',
' WHERE UPPER(Name) LIKE UPPER(''%[STR Name]%'') ')!
-- INSERT TYPES DATA RECORDS --
INSERT INTO Types
(Name, LinkField, LinkTable)
VALUES
('Weekly', 'PeriodTypeID', 'Forecasts') !
INSERT INTO Types
(Name, LinkField, LinkTable)
VALUES
('Monthly', 'PeriodTypeID', 'Forecasts') !
INSERT INTO Types
(Name, LinkField, LinkTable)
VALUES
('Annual', 'PeriodTypeID', 'Forecasts') !