Learning & Innovation Research Manager| Project Management Institute (PMI)Spain
Data quality and quantity is particularly important as we think about leveraging AI on projects. Considerations include the diversity and comprehensiveness of the data that is available to us.
Have you ever encountered unexpected challenges or pitfalls while using data in your projects? How did you navigate the situation and find a resolution? Saving Changes...
Solomon OtemaSenior Manager, Towers and Structures| IHS Nigeria Ltd (IHS Holding Ltd)Victoria Island, Lagos, Nigeria
While working with ChatGPT 4, I realized that in terms of the outputs generated, the quality of the data mattered more than the quantity of data being fed into the LLM. The more refined the input data was, the more I got outputs that were close to what I was expecting. My analysis of why this was the case showed me that for areas where the model did not receive significant training, the model was only able to generate good outputs based on the quality of the inputs. Inputs (especially files) that were big more or less crashed the model and it did not generate any outputs. Saving Changes...
Jorge EscalanteProject Manager/Educator| FHI360Tegucigalpa, Francisco Morazán, Honduras
Certainly, unexpected challenges with data quality and quantity are common in AI projects. One challenge I encountered involved incomplete and inconsistent data sources, leading to skewed model predictions. To navigate this, I employed data preprocessing techniques such as imputation and outlier removal to enhance data quality. Additionally, I collaborated closely with domain experts to gain insights into potential biases and limitations within the data. By iteratively refining data collection methods and ensuring diverse data sources, we improved the robustness and reliability of our AI models. This experience underscored the importance of thorough data validation and ongoing monitoring to mitigate unforeseen challenges and enhance the effectiveness of AI-driven solutions. Saving Changes...
Vanessa ThomasCybersecurity Project Manager, Data Quality, Change Management| AssystOdenton, Md, United States
Continuous monitoring data work flows, mange snterprise chnages made to tools used to gather and report data. Ensure compliance levels are met . Saving Changes...
Michael ShostSenior Security Portfolio Leader| Group 1001 SolutionsBrewster, NY, USA, United States
In my role spearheading a comprehensive data governance initiative for a state government agency, we encountered a significant challenge with the integrity, quality, and governance of the data available. The data landscape was fragmented, with disparate sources and inconsistent data management practices across different departments. This lack of cohesion posed considerable obstacles to leveraging AI effectively in various projects and initiatives.
One particular instance highlighted the critical need for addressing data quality and governance issues. We were tasked with implementing an enterprise data warehouse with AI driven components. However, upon closer examination of the available data, we discovered pervasive issues ranging from incomplete records and outdated information to privacy concerns and data silos.
To address these challenges, I led the development and implementation of a robust data governance framework encompassing data quality assurance, privacy protection, metadata management, and data sharing protocols. Working closely with stakeholders from diverse departments, we established standardized processes and guidelines to ensure data integrity and compliance with regulatory requirements.
As part of the program, we conducted comprehensive data audits and assessments to identify gaps and discrepancies in the existing data infrastructure. Leveraging industry best practices and frameworks such as DAMA (Data Management Association) and COBIT (Control Objectives for Information and Related Technologies), we devised strategies to improve data quality, enhance data lineage transparency, and establish clear ownership and accountability for data assets.
Furthermore, we deployed advanced data governance tools and platforms to automate data profiling, cleansing, and validation processes, thereby streamlining data management workflows and reducing manual errors. Through extensive training and capacity-building initiatives, we fostered a culture of data stewardship and collaboration among agency personnel, empowering them to take ownership of data quality and governance practices.
By proactively addressing the challenges related to data integrity, quality, and governance, we not only laid the foundation for successful AI implementations but also instilled confidence in the reliability and trustworthiness of the data assets underpinning critical government initiatives. This example underscores my leadership in driving transformative change through strategic data governance initiatives, ultimately enabling state agencies to harness the full potential of AI for public service delivery and decision-making. Saving Changes...
Access and quality of diverse data sets has often been a challenge in my past projects. When dealing with large scale datasets consisting of data which is secure and confidential at several levels it has always been crucial to balance access with data quality and security. Some questions which arose:
* Who has a need to access this data?
* Is it possible to meet that need without providing complete access?
* Is the limited dataset still of sufficient quality to be usable while still remaining secure? Saving Changes...
SHIBAJI BISWASPrincipal Project Engineer| CENTRE FOR RAILWAY INFORMATION SYSTEMSKOLKATA, WB, India
In my previous project, one of the objectives was to automatically fetch data from different applications managed by different project groups through API calls based on some user input data in my project's application. Due to the unavailability of data standardization, and uniformity of data format, the fetched data were wrong several times. So, we started a master data management project to develop a data governance framework. We worked with relevant stakeholders to define comprehensive data standards. This includes standards for data formats, structures, naming conventions, and units of measurement. we involved stakeholders from different departments in the standardization process and communicated the importance of data standardization across the organization. Saving Changes...
Anonymous
Emphasis is on data savviness, thus bringing in the concepts of data quality, monitoring and use of algorithms. Saving Changes...
Soraya KingU.S. Office of Personnel ManagementStafford, Va, United States
We were constantly going back to historical documents and files to verify data on projects. Unfortunately, most of the historical data was obsolete. So we acquired access to a program with regularly updated cost data and we work to stay abreast of updated industry standards. We also developed a data quality improvement plan to effectively manage data across the board. Saving Changes...
Wawan RidwanDriving project success through community and collaboration.| Merdeka Copper GoldJakarta, Indonesia
As a PMO Manager in a mining company in Indonesia, I've certainly faced challenges with data quality and quantity when applying AI to our projects. The key issues often revolve around the diversity and comprehensiveness of the data.
One common challenge is dealing with incomplete or inconsistent data, especially from remote locations where data collection can be sporadic. To overcome this, we've implemented stricter data collection protocols and employed IoT devices for more consistent data capture.
Another issue is ensuring data diversity, as AI models trained on limited data types can lead to biased results. We've addressed this by collaborating with other departments and external partners to broaden our data sources.
To ensure data comprehensiveness, we periodically review our data collection methods and the data types we gather. This helps identify gaps and make necessary adjustments.
Navigating these challenges requires a proactive approach, continuous monitoring, and adaptation of our data strategies to align with the evolving nature of our projects and AI advancements. Saving Changes...
In my current transformation initiative, the data challenges make up most of the business requirements. We have learned that one of the most important aspects is the structure of the data and the correct interpretation of the relationship between an asset and its attributes and contextual data. Saving Changes...
PMI Team Member
