diff options
| -rw-r--r-- | acm-data-economy22-3.txt | 39 | ||||
| -rw-r--r-- | acm-data-economy22-9.txt | 58 |
2 files changed, 97 insertions, 0 deletions
diff --git a/acm-data-economy22-3.txt b/acm-data-economy22-3.txt new file mode 100644 index 0000000..bdb86f9 --- /dev/null +++ b/acm-data-economy22-3.txt @@ -0,0 +1,39 @@ +Summary +------- + +This paper provides a systematic survey of existing data marketplaces and data +products therein. Specifically, the authors scraped information about 215,075 +data products from 15 data marketplaces. This information included metadata +such as price, description, data category, geographic scope, time range, etc. +Some of the findings obtained from analyzing these products include: +* perhaps surprisingly, most data products are offered for free, with only + 11,823 paid products. +* the paid products are sold mostly under either a subscription-based model + (for “live” data) or a one-off model. +* the range of data prices is very wide, but prices can often be well predicted + by simple regression model using features such as number of data points, + coverage and granularity. +* the relative predictability of prices can be used to build a data quotation + tool that takes as input the description of a new data product and outputs + a price quotation. + +Comments +-------- + +This works provides a valuable empirical study of existing data marketplaces. +The paper is well-written and the methodology is well-documented. The analysis +of the scraped data products is a convincing first step in trying to make sense +of the wide range of data products and the factors driving their prices. + +Minor comments: +* the second sentence of the first paragraph in section 6.1 seems to be the + result of an incomplete editing process and should be rewritten. Probably, + "We found that...". +* the y-axis in Figure 4 should be labeled “Number of products”. + +Conclusion +---------- + +I believe that this works is a valuable reference point and asset to the Data +Economy research community and I recommend acceptance. + diff --git a/acm-data-economy22-9.txt b/acm-data-economy22-9.txt new file mode 100644 index 0000000..fb32e0a --- /dev/null +++ b/acm-data-economy22-9.txt @@ -0,0 +1,58 @@ +Summary +------- + +This paper presents and studies data spaces, which can be thought of a specific +kind of data exchange/market. The defining feature of a data space, compared to +generic data marketplaces, appears to be a greater focus on interoperability, +allowing value to be extracted from data more easily, typically within +a specific application domain. + +A (fictional?) use case that is used as a running example throughout the paper +is “Green twin”: a data exchange pooling together data about a city's +buildings, vehicles, inhabitants and network infrastructure with the goal of +improving energy efficiency and quality of life. A survey of existing efforts +in the realm of data spaces (led by two related non-profit organizations, IDSA +and Gaia-X) is presented, highlighting the challenges of data interoperability +(with possible solutions involving the creation of data ontologies and the use +of standardized protocols) and generating value from data (with possible +solutions involving the use of automated machine learning techniques with +transfer of knowledge). + +Comments +-------- + +While the concept of data spaces seems promising and an interesting object of +study, this paper suffers from a somewhat ambiguous scope and its contributions +are not immediately apparent: + +1. As an overview of existing techniques and solutions, many of the + explanations were lacking in precision. For example, I was not able to find + a clear definition of data spaces and how they differ from data + marketplaces, and had to rely instead on slowly discovering the concept over + the course of the entire paper. The closest to this were the following + sentences in the introduction: + + This concept serves as an abstraction for data management in case where + many stakeholders are involved and exchange data with each other. The + easy data exchange between the stakeholders will generate value, + especially in combination with data analytics. New trading mechanisms + can allow stakeholders to cooperate with each other based on the value + of the exchanged data and the analytics services. + +2. As a position paper, I found that the description of future challenges lacks + in concreteness. What are the concrete open questions that researchers in + the data economy community should focus on? Do we need new algorithms? new + data structures? new machine learning techniques? What are specific ways in + which currently existing techniques unable to solve this challenges? + +Overall, I think the paper would benefit from focusing on at most two of the +following three kind of contributions: +1. A systematic description and documentation of currently existing + technologies, protocols, and standards that can be used to build a data + space. +2. A concrete proposal for the “Green twin” use case following the standards of + the systems community: a clear description of a system that would solve this + use case, the different pieces it will contain and how they interact with + other. +3. Concrete open questions and conjectures and an invitation to the research + community to work on them. |