You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[](https://www.repostatus.org/#active)
**reviser** is an R package designed for working with time-series vintages data. The package provides tools to clean, visualize, and analyze time-series revisions.
26
+
**reviser** is an R package designed for working with time-series
27
+
vintages data. The package provides tools to clean, visualize, and
28
+
analyze time-series revisions.
24
29
25
30
## Why reviser?
26
31
27
-
Economic data often arrives in multiple waves—with initial estimates updated or revised as more complete information becomes available (See the vignette [**the role and importance of revisions**](https://docs.ropensci.org/reviser/articles/literature-review.html)). These revisions, while common, can have major implications for how economic conditions are perceived and how decisions are made by policymakers, analysts, and markets. Yet, tools to systematically analyze, visualize, and communicate these revisions are still limited. This is where the R package **reviser** comes in.
28
-
29
-
**reviser** is built to support transparent, reproducible workflows for tracking and interpreting data revisions. Whether you're evaluating GDP estimates, inflation statistics, or high-frequency indicators, **reviser** helps quantify how vintages evolve, assess their reliability, and highlight cases where revisions may alter the economic narrative.
32
+
Economic data often arrives in multiple waves—with initial estimates
33
+
updated or revised as more complete information becomes available (See
These revisions, while common, can have major implications for how
37
+
economic conditions are perceived and how decisions are made by
38
+
policymakers, analysts, and markets. Yet, tools to systematically
39
+
analyze, visualize, and communicate these revisions are still limited.
40
+
This is where the R package **reviser** comes in.
41
+
42
+
**reviser** is built to support transparent, reproducible workflows for
43
+
tracking and interpreting data revisions. Whether you’re evaluating GDP
44
+
estimates, inflation statistics, or high-frequency indicators,
45
+
**reviser** helps quantify how vintages evolve, assess their
46
+
reliability, and highlight cases where revisions may alter the economic
47
+
narrative.
30
48
31
49
### Real-world relevance
32
-
The importance of revisions isn’t just academic. They shape real-world outcomes:
33
-
34
-
35
-
36
-
- US GDP 2015-Q1 --- Weather or Weakness?: A sharp downward revision from +0.1% to -0.2% sparked fears of a slowdown—until a later benchmark revision lifted it back to +0.8%, highlighting challenges in seasonal adjustment.
37
50
38
-
- Euro Area GDP in 2012Q1 --- A Recession Delayed?: A flash estimate of 0.0% avoided the recession label—until it was revised to -0.1%, confirming back-to-back contractions and altering the policy discussion at a critical time.
39
-
40
-
These examples underscore how even small numerical changes can shift narratives, delay responses, and affect credibility.
41
-
42
-
**reviser** is an R package designed to streamline the analysis and visualization of data revisions—especially in the context of official statistics and macroeconomic indicators. Built with tidy principles and seamless integration in mind, reviser offers intuitive tools to compare data vintages, quantify revision patterns, and produce publication-ready outputs. Whether you're tracking GDP estimate updates or evaluating forecast accuracy over time, **reviser** provides a robust and flexible framework tailored for economists, data analysts, and statistical agencies alike.
43
-
44
-
The **reviser** package provides a comprehensive toolkit for analyzing data revisions — crucial for anyone working with real-time data. It allows users to **visualize**, **analyze**, and **evaluate** the impact of data updates across different release vintages, helping to understand and analyze revision patterns.
51
+
The importance of revisions isn’t just academic. They shape real-world
52
+
outcomes:
53
+
54
+
<figure>
55
+
<img src="man/figures/gdp_over_time.jpeg"
56
+
alt="Revision paths for Euro Area GDP (Sources: BEA, Eurostat)" />
57
+
<figcaptionaria-hidden="true">Revision paths for Euro Area GDP
58
+
(Sources: BEA, Eurostat)</figcaption>
59
+
</figure>
60
+
61
+
- US GDP 2015-Q1 — Weather or Weakness?: A sharp downward revision from
62
+
+0.1% to -0.2% sparked fears of a slowdown—until a later benchmark
63
+
revision lifted it back to +0.8%, highlighting challenges in seasonal
64
+
adjustment.
65
+
66
+
- Euro Area GDP in 2012Q1 — A Recession Delayed?: A flash estimate of
67
+
0.0% avoided the recession label—until it was revised to -0.1%,
68
+
confirming back-to-back contractions and altering the policy
69
+
discussion at a critical time.
70
+
71
+
These examples underscore how even small numerical changes can shift
72
+
narratives, delay responses, and affect credibility.
73
+
74
+
**reviser** is an R package designed to streamline the analysis and
75
+
visualization of data revisions—especially in the context of official
76
+
statistics and macroeconomic indicators. Built with tidy principles and
77
+
seamless integration in mind, reviser offers intuitive tools to compare
78
+
data vintages, quantify revision patterns, and produce publication-ready
79
+
outputs. Whether you’re tracking GDP estimate updates or evaluating
80
+
forecast accuracy over time, **reviser** provides a robust and flexible
81
+
framework tailored for economists, data analysts, and statistical
82
+
agencies alike.
83
+
84
+
The **reviser** package provides a comprehensive toolkit for analyzing
85
+
data revisions — crucial for anyone working with real-time data. It
86
+
allows users to **visualize**, **analyze**, and **evaluate** the impact
87
+
of data updates across different release vintages, helping to understand
88
+
and analyze revision patterns.
45
89
46
90
Get started:
47
91
48
-
- Structure your data according to **reviser conventions**. See the [*get started*](https://docs.ropensci.org/reviser/articles/reviser.html) vignette
92
+
- Structure your data according to **reviser conventions**. See the
-**Calculate revisions** across vintages using [`get_revisions()`](https://docs.ropensci.org/reviser/reference/get_revisions.html). See the vignette [*Understanding Data Revisions*](https://docs.ropensci.org/reviser/articles/understanding-revisions.html) to learn how to structure and compute revision tables.
53
-
54
-
-**Analyze revision patterns** and evaluate revision accuracy and bias using [`get_revision_analysis()`](https://docs.ropensci.org/reviser/reference/get_revision_analysis.html). For more, read the vignette [*Revision Patterns and Statistics*](https://docs.ropensci.org/reviser/articles/revision-analysis.html).
55
-
56
-
-**Detect the first efficient release**, i.e., the earliest vintage that closely matches the final values, with [`get_first_efficient_release()`](https://docs.ropensci.org/reviser/reference/get_first_efficient_release.html). See the vignette [*Efficient Release Identification*](https://docs.ropensci.org/reviser/articles/efficient-release.html).
57
-
58
-
-**Nowcast future data revisions** using [`kk_nowcast()`](https://docs.ropensci.org/reviser/reference/kk_nowcast.html) and [`jvn_nowcast()`](https://docs.ropensci.org/reviser/reference/jvn_nowcast.html), tools to anticipate upcoming changes to early releases. Explore the methodology in the vignettes [*Nowcasting revisions using the generalized Kishor-Koenig family*](https://docs.ropensci.org/reviser/articles/nowcasting-revisions-kk.html) and [*Nowcasting revisions using the Jacobs-Van Norden model*](https://docs.ropensci.org/reviser/articles/nowcasting-revisions-jvn.html).
#> F-statistic: 3314 on 1 and 166 DF, p-value: < 2.2e-16
202
264
#>
203
265
#>
@@ -218,12 +280,29 @@ summary(efficient_release)
218
280
```
219
281
220
282
## Comparison to [`rjd3revisions`](https://rjdverse.github.io/rjd3revisions/)
221
-
The `reviser` package sets itself apart from `rjd3revisions` not only through its focus on advanced analysis of efficient releases and nowcasting performance, but also in its pure R implementation, which avoids external dependencies. In contrast, `rjd3revisions` relies heavily on Java via the JDemetra+ platform, which can make setup and integration more complex. `reviser` offers a lightweight, R-native solution for revision analysis, combining user-friendly tools for data wrangling, visualization, and evaluation of release efficiency.
283
+
284
+
The `reviser` package sets itself apart from `rjd3revisions` not only
285
+
through its focus on advanced analysis of efficient releases and
286
+
nowcasting performance, but also in its pure R implementation, which
287
+
avoids external dependencies. In contrast, `rjd3revisions` relies
288
+
heavily on Java via the JDemetra+ platform, which can make setup and
289
+
integration more complex. `reviser` offers a lightweight, R-native
290
+
solution for revision analysis, combining user-friendly tools for data
291
+
wrangling, visualization, and evaluation of release efficiency.
222
292
223
293
## Contributing
224
-
Contributions are encouraged and appreciated. If you're uncertain about opening a pull request, consider starting with an issue to discuss your proposal. For more information, please refer to [CONTRIBUTING](https://docs.ropensci.org/reviser/CONTRIBUTING.html).
225
294
226
-
Please note that the reviser project follows the [rOpenSci Code of Conduct](https://devguide.ropensci.org/collaboration.html#code-of-conduct). By contributing to this project, you agree to abide by its terms.
295
+
Contributions are encouraged and appreciated. If you’re uncertain about
296
+
opening a pull request, consider starting with an issue to discuss your
By contributing to this project, you agree to abide by its terms.
227
303
228
304
## Citation
229
-
Burri M, Wegmueller P (2025). reviser: Tools for Studying Revision Properties in Real-Time Time Series Vintages. R package version 0.1.0.9001, https://docs.ropensci.org/reviser/.
305
+
306
+
Burri M, Wegmueller P (2025). reviser: Tools for Studying Revision
307
+
Properties in Real-Time Time Series Vintages. R package version
0 commit comments