datarobot-predictions

Name: datarobot-predictions
Author: datarobot-oss

22stars19forks

How to install datarobot-predictions?

datarobot-oss/datarobot-agent-skills/datarobot-predictions

$npx -y skills add datarobot-oss/datarobot-agent-skills --skill datarobot-predictions

Installs into the current project.

Summary

Tools and guidance for making predictions with DataRobot deployments, including real-time predictions, batch scoring, prediction dataset generation, and prediction explanations (SHAP/XEMP). Use when making predictions, running batch scoring, generating prediction datasets, or exp

Files · 1

View on GitHub

SKILL.md

1	# DataRobot Predictions Skill
2
3	This skill provides comprehensive guidance for working with DataRobot predictions, including real-time predictions, batch scoring, and generating prediction datasets.
4
5	## Quick Start
6
7	Most common use case: Generate predictions for a deployment
8
9	1. Get deployment features: `get_deployment_features(deployment_id)` to understand required columns
10	2. Generate template: `generate_prediction_data_template(deployment_id, n_rows)` to create CSV structure
11	3. Make predictions: Use `deployment.predict_batch(...)` (works for both single-row “real-time” and batch scoring)
12
13	Example: "Generate a prediction dataset template for deployment abc123 with 10 rows"
14
15	To also explain predictions: pass `--max-explanations N` to `make_prediction.py` (or the
16	`max_explanations=N` kwarg in code). See [Prediction Explanations](#prediction-explanations) below.
17
18	## When to use this skill
19
20	Use this skill when you need to:
21	- Make predictions from deployed DataRobot models
22	- Explain individual predictions from a deployment (SHAP or XEMP, per-row)
23	- Generate prediction dataset templates
24	- Validate prediction data before scoring
25	- Understand deployment feature requirements
26	- Perform batch predictions on large datasets
27	- Get sample training data to understand expected formats
28
29	> For post-hoc explanations against a training project / leaderboard model (not a deployment),
30	> use the `datarobot-model-explainability` skill instead. This skill covers deployment-time
31	> explanations returned alongside scoring.
32
33	## Key capabilities
34
35	### 1. Understanding Deployment Requirements
36
37	Before making predictions, you need to understand what features a deployment requires:
38
39	- Feature names and types: Know which columns are needed (numeric, categorical, text, date)
40	- Feature importance: Understand which features matter most
41	- Target information: Know what you're predicting
42	- Time series configuration: If applicable, understand datetime columns and series IDs
43
44	### 2. Generating Prediction Datasets
45
46	Create properly formatted prediction datasets:
47
48	- Generate CSV templates with all required columns
49	- Include sample values appropriate for each feature type
50	- Add metadata comments explaining the model structure
51	- Ensure correct column ordering
52
53	### 3. Validating Prediction Data
54
55	Validate datasets before making predictions:
56
57	- Check for missing required features
58	- Verify data types match expected types
59	- Identify missing low-importance features (warnings)
60	- Note extra columns that will be ignored
61
62	### 4. Making Predictions
63
64	Execute predictions using various methods:
65
66	- Real-time predictions: Fast, synchronous predictions for individual records
67	- Batch predictions: Process large datasets efficiently
68	- Time series predictions: Handle forecasting scenarios with proper datetime handling
69
70	## Workflow examples
71
72	### Example 1: Generate prediction dataset for a new scenario
73
74	User request: "I want to predict sales for next week for store_A with temperatures of 75°F each day and no promotions."
75
76	Agent workflow:
77	1. Get deployment features to understand required columns
78	2. Generate a prediction data template with 7 rows (one week)
79	3. Fill in the template with user's specific values:
80	- Set temperature = 75 for all rows
81	- Set promotion = 0 for all rows
82	- Set store_id = "store_A" for all rows
83	- Set dates for next 7 days
84	4. Validate the data to ensure it's correct
85	5. Make predictions using the validated dataset
86
87	### Example 2: Batch scoring a CSV file
88
89	User request: "Score all records in my prediction_data.csv file using deployment abc123."
90
91	Agent workflow:
92	1. Validate the CSV file structure matches deployment requirements
93	2. Upload the file or provide file path
94	3. Submit batch prediction job
95	4. Monitor job status
96	5. Retrieve and return prediction results
97
98	## Using DataRobot SDK
99
100	This skill guides you to use the DataRobot Python SDK directly. Install the SDK if needed:
101
102	```bash
103	pip install datarobot
104	```
105
106	### Key SDK Operations
107
108	Use these DataRobot SDK methods to work with predictions:
109
110	Deployment Information:
111	- `dr.Deployment.get(deployment_id)` - Get deployment details
112	- `deployment.get_features()` - Get required features (name/type/importance)
113
114	Predictions:
115	- `deployment.predict_batch(source)` - Convenience batch prediction API (CSV path, file object, or pandas DataFrame)
116	- `dr.BatchPredictionJob.score(deployment=deployment, ...)` - Advanced batch prediction control
117	- `job.get_result_when_complete()` - Wait for batch scoring to finish and download results
118
119	Data Management:
120	- `dr.Dataset.c

People who liked this also viewed

6 picks

Type

datarobot-predictions

22stars19forks

View on GitHub

How to install datarobot-predictions?

datarobot-oss/datarobot-agent-skills/datarobot-predictions

$npx -y skills add datarobot-oss/datarobot-agent-skills --skill datarobot-predictions

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# DataRobot Predictions Skill
2
3	This skill provides comprehensive guidance for working with DataRobot predictions, including real-time predictions, batch scoring, and generating prediction datasets.
4
5	## Quick Start
6
7	Most common use case: Generate predictions for a deployment
8
9	1. Get deployment features: `get_deployment_features(deployment_id)` to understand required columns
10	2. Generate template: `generate_prediction_data_template(deployment_id, n_rows)` to create CSV structure
11	3. Make predictions: Use `deployment.predict_batch(...)` (works for both single-row “real-time” and batch scoring)
12
13	Example: "Generate a prediction dataset template for deployment abc123 with 10 rows"
14
15	To also explain predictions: pass `--max-explanations N` to `make_prediction.py` (or the
16	`max_explanations=N` kwarg in code). See [Prediction Explanations](#prediction-explanations) below.
17
18	## When to use this skill
19
20	Use this skill when you need to:
21	- Make predictions from deployed DataRobot models
22	- Explain individual predictions from a deployment (SHAP or XEMP, per-row)
23	- Generate prediction dataset templates
24	- Validate prediction data before scoring
25	- Understand deployment feature requirements
26	- Perform batch predictions on large datasets
27	- Get sample training data to understand expected formats
28
29	> For post-hoc explanations against a training project / leaderboard model (not a deployment),
30	> use the `datarobot-model-explainability` skill instead. This skill covers deployment-time
31	> explanations returned alongside scoring.
32
33	## Key capabilities
34
35	### 1. Understanding Deployment Requirements
36
37	Before making predictions, you need to understand what features a deployment requires:
38
39	- Feature names and types: Know which columns are needed (numeric, categorical, text, date)
40	- Feature importance: Understand which features matter most
41	- Target information: Know what you're predicting
42	- Time series configuration: If applicable, understand datetime columns and series IDs
43
44	### 2. Generating Prediction Datasets
45
46	Create properly formatted prediction datasets:
47
48	- Generate CSV templates with all required columns
49	- Include sample values appropriate for each feature type
50	- Add metadata comments explaining the model structure
51	- Ensure correct column ordering
52
53	### 3. Validating Prediction Data
54
55	Validate datasets before making predictions:
56
57	- Check for missing required features
58	- Verify data types match expected types
59	- Identify missing low-importance features (warnings)
60	- Note extra columns that will be ignored
61
62	### 4. Making Predictions
63
64	Execute predictions using various methods:
65
66	- Real-time predictions: Fast, synchronous predictions for individual records
67	- Batch predictions: Process large datasets efficiently
68	- Time series predictions: Handle forecasting scenarios with proper datetime handling
69
70	## Workflow examples
71
72	### Example 1: Generate prediction dataset for a new scenario
73
74	User request: "I want to predict sales for next week for store_A with temperatures of 75°F each day and no promotions."
75
76	Agent workflow:
77	1. Get deployment features to understand required columns
78	2. Generate a prediction data template with 7 rows (one week)
79	3. Fill in the template with user's specific values:
80	- Set temperature = 75 for all rows
81	- Set promotion = 0 for all rows
82	- Set store_id = "store_A" for all rows
83	- Set dates for next 7 days
84	4. Validate the data to ensure it's correct
85	5. Make predictions using the validated dataset
86
87	### Example 2: Batch scoring a CSV file
88
89	User request: "Score all records in my prediction_data.csv file using deployment abc123."
90
91	Agent workflow:
92	1. Validate the CSV file structure matches deployment requirements
93	2. Upload the file or provide file path
94	3. Submit batch prediction job
95	4. Monitor job status
96	5. Retrieve and return prediction results
97
98	## Using DataRobot SDK
99
100	This skill guides you to use the DataRobot Python SDK directly. Install the SDK if needed:
101
102	```bash
103	pip install datarobot
104	```
105
106	### Key SDK Operations
107
108	Use these DataRobot SDK methods to work with predictions:
109
110	Deployment Information:
111	- `dr.Deployment.get(deployment_id)` - Get deployment details
112	- `deployment.get_features()` - Get required features (name/type/importance)
113
114	Predictions:
115	- `deployment.predict_batch(source)` - Convenience batch prediction API (CSV path, file object, or pandas DataFrame)
116	- `dr.BatchPredictionJob.score(deployment=deployment, ...)` - Advanced batch prediction control
117	- `job.get_result_when_complete()` - Wait for batch scoring to finish and download results
118
119	Data Management:
120	- `dr.Dataset.c

People who liked this also viewed

6 picks

Type