Scenarios

import pandas as pd

dataframe = pd.read_csv("data/example.csv")
dataframe

	ID	Date	Continent	Country	City	Color	Quantity	Price
0	1	2019-01-01	Europe	France	Paris	red	1000.0	500.0
1	2	2019-01-02	Europe	France	Lyon	red	2000.0	400.0
2	3	2019-01-05	Europe	France	Paris	blue	3000.0	420.0
3	4	2018-01-01	Europe	France	Bordeaux	blue	1500.0	480.0
4	5	2019-01-01	Europe	UK	London	green	3000.0	460.0
5	6	2019-01-01	Europe	UK	London	red	2500.0	500.0
6	7	2019-01-02	Asia	China	Beijing	blue	2000.0	410.0
7	8	2019-01-05	Asia	China	HongKong	green	4000.0	350.0
8	9	2018-01-01	Asia	India	Dehli	red	2200.0	360.0
9	10	2019-01-01	Asia	India	Mumbai	blue	1500.0	400.0

import atoti as tt

session = tt.create_session()
store = session.read_pandas(dataframe, "First", keys=["ID"])
cube = session.create_cube(store, "FirstCube")
cube.query()

Welcome to atoti 0.4.0!

By using this community edition, you agree with the license available at https://www.atoti.io/eula.
Browse the official documentation at https://docs.atoti.io.
Join the community at https://www.atoti.io/register.

You can hide this message by setting the ATOTI_HIDE_EULA_MESSAGE environment variable to True.

	Price.MEAN	Price.SUM	Quantity.MEAN	Quantity.SUM	contributors.COUNT
0	428.0	4280.0	2270.0	22700.0	10

Scenarios from modified sources

Load a modified version of your data into the store to compare it to the original version.

First, let’s modify the source dataset with pandas:

dataframe.loc[(dataframe["City"] == "Paris"), "Color"] = "purple"
dataframe

	ID	Date	Continent	Country	City	Color	Quantity	Price
0	1	2019-01-01	Europe	France	Paris	purple	1000.0	500.0
1	2	2019-01-02	Europe	France	Lyon	red	2000.0	400.0
2	3	2019-01-05	Europe	France	Paris	purple	3000.0	420.0
3	4	2018-01-01	Europe	France	Bordeaux	blue	1500.0	480.0
4	5	2019-01-01	Europe	UK	London	green	3000.0	460.0
5	6	2019-01-01	Europe	UK	London	red	2500.0	500.0
6	7	2019-01-02	Asia	China	Beijing	blue	2000.0	410.0
7	8	2019-01-05	Asia	China	HongKong	green	4000.0	350.0
8	9	2018-01-01	Asia	India	Dehli	red	2200.0	360.0
9	10	2019-01-01	Asia	India	Mumbai	blue	1500.0	400.0

We can now load this dataset into a new scenario:

store.scenarios["Purple Paris"].load_pandas(dataframe)
store.scenarios["Purple Paris"]

• First
  • ID
    • key: True
    • nullable: False
    • type: long
  • Date
    • key: False
    • nullable: True
    • type: LocalDate[yyyy-MM-dd]
  • Continent
    • key: False
    • nullable: True
    • type: string
  • Country
    • key: False
    • nullable: True
    • type: string
  • City
    • key: False
    • nullable: True
    • type: string
  • Color
    • key: False
    • nullable: True
    • type: string
  • Quantity
    • key: False
    • nullable: False
    • type: double
  • Price
    • key: False
    • nullable: False
    • type: double

If you want the store to appear in application’s Source Simulation widget, you will need to enable this parameter on the store

store.source_simulation_enabled = True

Scenarios measure scaling

You can also perform simulations by modifying the value of certain measures. There are three available simulation methods:

• multiply: multiplies the measure’s value with the provided weight

• replace: replaces the measure’s value with the given value

• add: increments the measure’s value with the provided value

help(cube.setup_simulation)

Help on method setup_simulation in module atoti.cube:

setup_simulation(name: 'str', *, base_scenario: 'str' = 'Base', levels: 'Optional[Sequence[Level]]' = None, multiply: 'Optional[Collection[Measure]]' = None, replace: 'Optional[Collection[Measure]]' = None, add: 'Optional[Collection[Measure]]' = None) -> 'Simulation' method of atoti.cube.Cube instance
    Create a simulation store for the given measures.

    Simulations can have as many scenarios as desired.

    The same measure cannot be passed in several methods.

    Args:
        name: The name of the simulation.
        base_scenario: The name of the base scenario.
        levels: The levels to simulate on.
        multiply: Measures whose values will be multiplied.
        replace: Measures whose values will be replaced.
        add: Measures whose values will be added (incremented).

    Returns:
        The simulation on which scenarios can be made.

lvl = cube.levels
m = cube.measures

Calling the cube.setup_simulation method will create a special type of store which controls the simulation. This store’s key fields are the levels you provide, as well as the column with the name of the simulation, which contains the scenario names.

simulation = cube.setup_simulation(
    "First simulation",
    levels=[lvl["Continent"]],
    multiply=[m["Quantity.SUM"], m["Quantity.MEAN"]],
    replace=[m["Price.MEAN"]],
)
simulation

• First simulation
  • Levels
    1. Continent
  • Scenarios
    1. Base
  • multiply
    1. Quantity.SUM
    2. Quantity.MEAN
  • replace
    1. Price.MEAN

simulation.head()

		First simulation_Quantity.SUM_multiply	First simulation_Quantity.MEAN_multiply	First simulation_Price.MEAN_replace	Priority
Continent	First simulation

In order to populate this simulation with scenarios, you can either populate the simulation the same way you would a regular store, with a CSV file or a DataFrame. Or you can create a scenario for this simulation and then insert rows into it either manually or from a source file.

When filling a scenario, (not the simulation), there are several things which are done automatically. For instance, you don’t need to specify the priority.

asian_growth = simulation.scenarios["Asian growth"]
asian_growth += ("Asia", 1.05, 1.05, 20)
asian_growth += ("Europe", 0.9, 0.9, 0.85)
asian_growth.head()

	First simulation_Quantity.SUM_multiply	First simulation_Quantity.MEAN_multiply	First simulation_Price.MEAN_replace	Priority
Continent
Asia	1.05	1.05	20.00	0.0
Europe	0.90	0.90	0.85	0.0

Calling head on the scenario will display the rows of the Simulation handled by this scenario.

cube.query(
    m["Quantity.SUM"],
    m["Quantity.MEAN"],
    m["Price.MEAN"],
    levels=[lvl["Continent"], lvl["First simulation"]],
)

		Quantity.SUM	Quantity.MEAN	Price.MEAN
Continent	First simulation
Asia	Base	9700.0	2425.000000	380.00
	Asian growth	10185.0	2546.250000	20.00
Europe	Base	13000.0	2166.666667	460.00
	Asian growth	11700.0	1950.000000	0.85

You can create as many scenarios as you want. However, every scenario of a simulation uses the same methods. If you want to use a different method you must create a new simulation.

When loading a DataFrame into the simulation, you can get the required headers from the simulation. The same is true for a scenario. For the priority, you can use any numeric value you want. When loading a DataFrame into a scenario, the priority column is optional.

growth = simulation.scenarios["Growth"]
normal_priority = 0
rows = [
    ["Asia", 1.1, 1.1, 22, normal_priority],
    ["Europe", 1.2, 1.3, 15, normal_priority],
]
df = pd.DataFrame(rows, columns=growth.columns)
df

	Continent	First simulation_Quantity.SUM_multiply	First simulation_Quantity.MEAN_multiply	First simulation_Price.MEAN_replace	Priority
0	Asia	1.1	1.1	22	0
1	Europe	1.2	1.3	15	0

growth = simulation.scenarios["Growth"].load_pandas(df)
growth.head()

	First simulation_Quantity.SUM_multiply	First simulation_Quantity.MEAN_multiply	First simulation_Price.MEAN_replace	Priority
Continent
Asia	1.1	1.1	22.0	0.0
Europe	1.2	1.3	15.0	0.0

You can delete a simulation from the cube’s simulation dict

del cube.simulations["First simulation"]

You can create simulations wihout any levels. These simulations can only have one value per scenario. For simulations whithout any fields and acting on only one measure, you can assign values to scenarios instead of inserting tuples.

no_field_sim = cube.setup_simulation("No Field Sim", multiply=[m["Price.SUM"]])
no_field_sim.scenarios["- 10 %"] = 0.9
no_field_sim.scenarios["+ 20%"] = 1.2
no_field_sim.head()

	No Field Sim_Price.SUM_multiply	Priority
No Field Sim
- 10 %	0.9	0.0
+ 20%	1.2	0.0

The priority rules are used to supersed the default behaviour between conflicting rules in the simulation. The basic rule is the more wildcard fields there are and the further to the left these wildcards are, the lower the priority. So in the following example, the rule on Europe overrides the one on France. To use wildcards, simply pass None as the value for the column(s) you want the wildcard to be on.

priority_sim = cube.setup_simulation(
    "Priority Simulation",
    levels=[lvl["Continent"], lvl["Country"]],
    multiply=[m["Quantity.MEAN"]],
)
priority_scenario = priority_sim.scenarios["growth"]
priority_scenario += ("Europe", None, 1.1)
priority_scenario += (None, "France", 1.2)
priority_scenario.head()

		Priority Simulation_Quantity.MEAN_multiply	Priority
Continent	Country
Europe	NaN	1.1	0.0
NaN	France	1.2	0.0

cube.query(
    m["Quantity.MEAN"],
    levels=[lvl["Continent"], lvl["Country"], lvl["Priority Simulation"]],
)

			Quantity.MEAN
Continent	Country	Priority Simulation
Asia	China	Base	3000.0
		growth	3000.0
	India	Base	1850.0
		growth	1850.0
Europe	France	Base	1875.0
		growth	2062.5
	UK	Base	2750.0
		growth	3025.0

To override the Europe rule with one specific to France, we can give a higher priority:

priority_scenario += (
    None,
    "France",
    120,
    normal_priority + 1,
)
priority_scenario.head()

		Priority Simulation_Quantity.MEAN_multiply	Priority
Continent	Country
Europe	NaN	1.1	0.0
NaN	France	120.0	1.0

cube.query(
    m["Quantity.MEAN"],
    levels=[lvl["Continent"], lvl["Country"], lvl["Priority Simulation"]],
)

			Quantity.MEAN
Continent	Country	Priority Simulation
Asia	China	Base	3000.0
		growth	3000.0
	India	Base	1850.0
		growth	1850.0
Europe	France	Base	1875.0
		growth	225000.0
	UK	Base	2750.0
		growth	3025.0